Organizers : Cheng Shiu University (正修科技大學) University of Hyogo (兵庫県立大学) Venue : Administration Building

Organizers : Cheng Shiu University (正修科技大學) University of Hyogo (兵庫県立大学)

Sponsors: Graduate School of Engineering, University of Hyogo (兵庫県立大学)＊ Nobuhara Hospital and Institute of Biomechanics (信原病院) This symposium obtains the approval of Himeji Medical Association.

Venue : Administration Building 12F, International Conference Hall, Cheng Shiu University, Kaohsiung, Taiwan

The 6th International Symposium in Computational Medical and Health Technology

Organizers：Cheng Shiu University (正修科技大學)

University of Hyogo (兵庫県立大学)

Sponsors：Graduate School of Engineering, University of Hyogo (兵庫県立大学)*

Nobuhara Hospital and Institute of Biomechanics (信原病院)

＊This symposium obtains the approval of Himeji Medical Association.

Host：Cheng Shiu University (正修科技大學)

Date： September 5-7, 2016

Venue：Administration Building 12F, International Conference

Hall, Cheng Shiu University, Kaohsiung, Taiwan

Program Manual

Content

I. Organization ...... 1

II. Schedule ...... 2

III. Venue ...... 7

IV. Invited Session 1 ...... 8

V. General Session 1 ...... 10

VI. Invited Session 2 ...... 14

VII. Invited Session 3 ...... 19

VIII. Invited Session 4 ...... 22

IX. Invited Session 5 ...... 26

X. General Session 2 ...... 31

XI. General Session 3 ...... 35

XII. Contact ...... 39

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016 I. Organization

Symposium Chair：Naotake Kamiura (University of Hyogo) Shuenn-Ren Cheng (Cheng Shiu University)

Vise Symposium Chair：Syoji Kobashi (University of Hyogo) Wu-Te Ko (Cheng Shiu University)

Honorary Chair: Katsuya Nobuhara(Nobuhara Hospital) Yutaka Hata (University of Hyogo)

Financial Chair：Masakazu Morimoto (University of Hyogo) Chao-Cheng Wu (Taipei Tech)

Program Chair：Hiroaki Inui (Nobuhara Hospital) Nanabu Nii (University of Hyogo)

Local Arrangement Chair：Takayuki Fujita (University of Hyogo) Juei-Chao Chen (Fu Jen University) Chao-Cheng Wu (Taipei Tech)

Publication Chair：Masaru Higa (University of Hyogo) Nozomu Araki (University of Hyogo)

Secretary: Hiroshi Tanaka (Nobuhara Hospital)

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

II. Schedule

Sunday September 4, 2016 Welcome Dinner Host by Dr. Jinne-Jie Wu 18:00 海天下海產餐廳高雄市新興區林森二路 188 號 TEL: 07-281-0651

Monday September 5, 2016 Opening 09:00 Shuenn-Ren Cheng, VICE PRESIDENT Cheng Shiu University, Kaohsiung, Taiwan Yutaka Hata, HONORARY CHAIR University of Hyogo, Himeji, Japan Invited Session 1 (Page 8) Yutaka Hata (University of Hyogo) Jih-Yang Ko (Kaohsiung Chang Gung Memorial Hospital) 09:10-09:50 Analysis of Throwing Motion: Rotator Interval Lesion and APIT Lesion Katsuya Nobuhara

09:50-10:30 Digital Healthcare – Vital Sign and Measurement Frequency Hiroshi Nakajima Break (10:30 – 10:40) General Session 1 (Page 10) Naomi Yagi (Hyogo College of Medicine) Toyohiko Hayashi (Niigata University) 10:40-11:00 Mechanisms of Shoulder Dislocation during Rugby Tackling Takayuki Kawasaki, Hiroshi Tanaka, Kenji Murakami, Toru Okuwaki, and Kazuo Kaneko

11:00-11:20 Vector Representations of Sentences Using the Word2vec for Nursing- Care Text Classification Yuya Tsuchida, Manabu Nii, Takuya Iwamoto, Atsuko Uchinuno, Reiko Sakashita

11:20-11:40 A Movement Analysis of Jaws and a Tongue by Use of EMG Signals for Chewing and Swallowing Evaluation Shota Okajima, Manabu Nii, Yuya Tsuchida, Takuya Iwamoto, Reiko Sakashita, Misao Hamada

11:40-12:00 Adaptive Array Processing for Noncontact Measurement of Heartbeat Using Microwaves Takuya Sakamoto, Kosuke Yamashita Lunch (12:00 – 13:30)

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Invited Session 2 (Page 14) Shyoji Kobashi (University of Hyogo) Shinich Yoshiya (Hyogo College of Medicine) 13:30-14:10 Current Advantages of Computer Navigation Versus Conventional Total Knee Arthroplasty in Alignment and Less Invasion: A Prospective Comparative Study on Endothelial Damage Marker Levels Jih-Yang Ko, Shu-Jui Kuo, Feng-Sheng Wang, Ching-Jen Wang, Wen-Yi Chou, Ka-Kit Siu

14:10-14:50 Preoperative Templating and Computer-Assisted Total Knee Arthroplasty for Arthritic Valgus Knee Wen-Yi Chou, Jih-Yang Ko, Ka-Kit Siu

14:50-15:30 The Application of Navigation Assisted Total Knee Arthroplasty for Osteoarthritis Knees with Extra-Articular Deformity Ka-Kit Siu, Ghe-Wu, Ching-Jen Wang, Wen-Yi Chow, Jih-Yang Koa Coffee break (15:30 – 15:50) Invited Session 3 (Page 19) Hiroshi Nakajima (OMRON Corporation) Hiroaki Inui (Nobuhara Hospital and Institute of Biomechanics) 15:50-16:30 Use of Computer Technology in Orthopaedic Surgery and Evaluation of the Surgical Outcome Shinichi Yoshiya

16:30-17:10 Research and Development in Collaboration With Private Sectors in Biomedical Engineering and Assistive Technology Toyohiko Hayashi, Hiroshi Tanaka, Hiroaki Inui, Masahiko Komai, Katsuya Nobuhara

17:10-17:50 MAKOPlasty Experince of 107 Cases (Robotic Arm Partial Knee Resurfacing) Juinn-Jer Wu, Li-Wei Wu, Ching-Chieh Yang Banquet Host by Prof. Shuenn-Ren Cheng 18:30 三集家常菜高雄市鳳山區文衡路 379 號 TEL:07-7777988

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Tuesday September 6, 2016 Invited Session 4 (Page 22) Takayuki Fujita (University of Hyogo) Toyohiko Hayashi (Niigata University) 09:30-10:10 Toward Automated Medical Practice Capsule Yutaka Hata

10:10-10:50 Understanding elderly’s personal differences in participating health- promotion exercise through motion data collection and analysis Tien-Lung Sun

10:50-11:30 The Implementation of Data Visualization Applications in Open-Source Software with Health IT Dashboard Utility Juei-Chao Chen, Chieh-Hsuan Liu, and Shuenn-Ren Cheng Lunch (11:30 – 13:00)

Invited Session 5 (Page 26) Naotake Kamiura (University of Hyogo) Hiroaki Inui (Nobuhara Hospital and Institute of Biomechanics) 13:00-13:40 Stress Analysis in Various Shoulder Surgeries by Finite Element Analysis Ching-Chieh Yang, Juinn-Jer Wu

13:40-14:20 Smartphone-based Statistical Gait Analysis Chao-Cheng Wu,

14:20-15:00 The Innovation of Orthopaedic Surgery by the Development of In Vivo 3D Kinematic Analysis System Kazuomi Sugamoto Coffee break (15:00 – 15:20) General Session 2 (Page 31) Masakazu Morimoto (University of Hyogo) Hiroshi Tanaka (Nobuhara Hospital and Institute Biomechanics) 15:20-15:40 Interconnection between Stretchable Wiring and Integrated Circuit Using Magnetic Attractive Force Motoki Tsukiyama, Shinya Toda, Ryosuke Fujita, Kensuke Kanda, Takayuki Fujita, and Kazusuke Maenaka

15:40-16:00 Wearable Pulse Sensing System with P(VDF/TrFE) Piezoelectric Film for Health Management Ryosuke Fujita, Shinya Toda, Motoki Tsukiyama, Kensuke Kanda, Takayuki Fujita, and Kazusuke Maenaka

16:00-16:20 Biomechanical Approach to the Shoulder Complex System Hiroshi Tanaka, Toyohiko Hayashi, Hiroaki Inui, Shyoji Kobashi, Shinich Yoshiya, Koji Shibano, Katsuya Nobuhara

16:20-16:40 Glenohumeral Relationships at Different Angles of Shoulder Abduction Hiroaki Inui, Hiroshi Tanaka, Katsuya Nobuhara Break (16:40 – 16:50)

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

General Session 3 (Page 35) Manabu Nii (University of Hyogo) Kazuomi Sugamoto (Orthopaedic Biomaterial Science) 16:50-17:10 Energy harvesting in IoT sensors Takayuki Fujita

17:10-17:30 Post-operative Implanted Knee Kinematics Prediction for Personalized Medicine Shoji Kobashi, Manabu Nii, Belayat Hossain, Syunichiro Kambara, Takatoshi Morooka, Makiko Okuno, and Shinichi Yoshiya

17:30-17:50 On Personal Identification Using Clinical Data Obtained by Optical Coherence Tomography Naotake Kamiura, Shoji Kobashi, Manabu Nii, Takayuki Yumoto, Tomofusa Yamauchi, and Hitoshi Tabuchi Closing 17:50-18:10 Naotake Kamiura, SYMPOSIUM CHAIR University of Hyogo, Himeji, Japan

Katsuya Nobuhara, HONORARY CHAIR Nobuhara Hospital and Institute of Biomechanics, Tatsuno, Japan Banquet Host by Prof. Chao-Cheng Wu 18:30 極鮮火鍋-鳳山澄清店高雄市三民區澄清路 708 號 TEL:07-7777956

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Wednesday September 7, 2016 Social Events 09:00-18:00 Social Events Lunch 12:00-13:00 味香山海產店高雄市大樹區統嶺路 72-3 號 TEL:07-6563608 Dinner Host by Dr. Katsuya Nobuhara 18:00 東港阿榮坊風味海鮮屏東縣光復路一段 399 號 TEL:08-833-3931

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

III. Venue

Main Gate

International Conference Hall (12F)

NO.840, CHENGCING RD., NIAOSONG DISTRIT, KAOHSIUNG CITY 833, TAIWAN(R.O.C.)

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

IV. Invited Session 1 (9:10-10:30, Sept. 5)

Analysis of Throwing Motion -Rotator Interval Lesion and APIT Lesion-

Katsuya Nobuhara, MD, PhD Nobuhara Hospital and Institute of Biomechanics, Tatsuno, Japan

ABSTRACT Throwing motion is a whole body exercise, not only by the upper arm but involves the movement of whole body. Maximum efficiency can be obtained when the legs are balanced, the power is accumulated by winding up ipsi-lateral arm and leg, and the fulcrum is transferred to contralateral leg. In other word, high speed propeller movement is attained. Throwing disorders mainly occur in the base on instability, and break out with overuse and repeated stress, and then, come out with the imbalance of the whole body. Clinical observation and biomechanical analysis on the dynamic activity of athletes have identified both organic and functional factors associated with its pathogenesis. Utilizing the motion analysis system, we are able to diagnose pathologic lesions caused by sports, to treat their lesions and prevent for sports disorder, and to guide athletes in their technical improvements. Rotator Interval lesion (RIL lesion), and Antero-Posterior Instability in Throwing plane (APIT lesion) will be discussed. Information about glenohumeral relationships at different angles of abduction is scarce and it remains unclear how angle of arm abduction affects shoulder function. The objective of the study was to clarify relationships between anatomical landmarks of the glenohumeral joint at different angles of abduction.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Digital Healthcare – Relationship between Vital Sign and Measurement Frequency

Hiroshi Nakajima, PhD Intelligence System Research Center, Technology and Intellectual Property, H.Q., Omron Corporation, Kyoto, Japan

ABSTRACT Lifestyle related diseases such as hypertension and diabetes are the main causes of long term nursing care. One of the most effective and efficient treatment and prevention of the diseases is individualized lifestyle modification based on the scientific point of view. The idea of Systems Healthcare has been developed based on data analysis using both field studies and internet services to get the new findings such as the seasonal, weekly, and circadian vital sign variability. They would provide important and effective information and knowledge in medical treatment and self-care. In this talk, the other important aspects of measurement are employed to estimate mind sets for health management. Because the data include time stamps individually, measurement frequency and time variability can be used in data analysis. Measurement frequency is used as the index of stratification analysis to prove the relationship between measurement behavior and vital signs such as blood pressure and weight. As the results, both vitals have some important tendencies according to the measurement frequency.

Keywords: digital healthcare, lifestyle change, vital and lifestyle monitoring

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

V. General Session 1 (10:40-12:00, Sept. 5)

Mechanisms of Shoulder Dislocation During Rugby Tackling

Takayuki Kawasaki1, MD, PhD, Hiroshi Tanaka2, PhD, Kenji Murakami3, PhD, Toru Okuwaki4, and Kazuo Kaneko1, MD, PhD 1Department of Orthopaedic surgery, Faculty of Medicine, Juntendo University, Tokyo, Japan 2Nobuhara Hospital and Institute of Biomechanics, Tatsuno, Japan, 3Sendai University, Faculty of Sports Science, Miyagi, Japan 4Japan Institute of Sports Science, Tokyo, Japan

ABSTRACT Our group especially aims to address characteristics of the shoulder and neck injury in collision athletes from diversified perspectives. The following projects are in progress: (1) Injury surveillance study. (2) Imaging analysis to illuminate the characteristics of collision athletes. (3) Video analysis to hypothesize mechanisms of the injuries. (4) Kinematic motion analysis to reproduce the shoulder joint and the neck at risk under experimental setting. (5) Model analysis to simulate kinetic aspects of these injuries. All of these studies may contribute to suggest proper treatment and indicate the preventive strategy for the injury. In this paper, we will mainly present our efforts to analyze the kinematic aspects of the tackling motion in 1-on-1 dual using a 3D motion capture system. These experimental setting was designed for the purpose of finding any key(s) to avoid traumatic anterior shoulder dislocation in rugby.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Vector Representations of Sentences Using the Word2vec for Nursing-Care Text Classification

Yuya Tsuchida1, Manabu Nii1, Takuya Iwamoto1, Atsuko Uchinuno2, Reiko Sakashita2 1Graduate School of Engineering, University of Hyogo, Japan 2College of Nursing Art and Science, University of Hyogo, Japan

ABSTRACT The number of staffs related to the medical field is insufficient because the number of work forces is decreasing. Therefore, we need a scheme to keep or improve current nursing quality under the small number of work forces. We have tackled nursing-care texts classification to improve nursing quality in several years. Nursing-care texts are free style Japanese texts. To classify nursing-care texts, we have used several machine learning methods. We used a word list, which consists of words appeared in the training data, for training a classification system. The relation among the words is not considered in our previous works. The word2vec is a tool to realize word representations in arbitrary dimensional space. Since the word vector representation enables computing the degree of similarity between words, we use the word2vec for representing relations between words. We propose a new feature vector definition using word vector representations realized by the word2vec. And we adopt multi-class SVMs to classify nursing-care texts. Our proposed method has a possibility to improve the classification performance.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

A Movement Analysis of Jaws and A Tongue by Use of EMG Signals for Chewing and Swallowing Evaluation

Shota Okajima1, Manabu Nii1, Takuya Iwamoto1, Yuya Tsuchida1, Reiko Sakashita2, Misao Hamada3 1Graduate School of Engineering, University of Hyogo, Japan 2College of Nursing Art and Science, University of Hyogo, Japan 3Social Welfare Corporation Lavita, Japan

ABSTRACT Chewing and swallowing ability is important to have a meal. Especially, chewing is known to improve oral and brain function. Nurses aim to give a meal corresponding to chewing and swallowing ability of patients. Chewing and swallowing ability of the patients is assessed by nurses with much experience of listening speciﬁc pronunciations. However, every nurse does not have the same assessment criteria of chewing and swallowing ability. Therefore, it is necessary for every nurse to know the movement of patients’ jaws and a tongue in order to assess chewing and swallowing ability of the patient. To obtain the movement of jaws and a tongue, we observe the surface EMG signals around jaws. Obtained EMG signals of some pronunciations are used to classify tongue movements using Deep Learning techniques. From some experiments, we can classify the surface EMG signals under large tongue movements with 80% of accuracy. In cases of small movement classification, we have 60% of accuracy.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Adaptive Array Processing for Noncontact Measurement of Heartbeat using Microwaves

Takuya Sakamoto Graduate School of Engineering, University of Hyogo, Japan

ABSTRACT Recent developments in mobile devices such as the Apple Watch have triggered consumer interest in monitoring vital signs in casual and non-medical settings. Such trends promote the development of technologies that can measure vital signs remotely without any contact sensors. Noncontact vital monitoring technologies enables the nonintrusive monitoring of physical and mental conditions in various applications. However, it is challenging to collect the desired signal containing only the vital signs, while rejecting other interfering signals. In this study, we apply adaptive array processing techniques to microwave echo data to retrieve heartbeat signals from an antenna array with multiple channels. The proposed method enhances the signal- to-noise power ratio, while rejecting unnecessary echoes by adaptively forming a sub-optimal array beam pattern. Then, the adaptively synthesized signal is processed using a topology algorithm to estimate an accurate instantaneous heart rate time sequence. We investigate the accuracy of the proposed method using a contact electrocardiography sensor as a reference.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

VI. Invited Session 2 (13:30-15:30, Sept. 5)

Current Advantages of Computer Navigation Versus Conventional Total Knee Arthroplasty in Alignment and Less Invasion: A Prospective Comparative Study on Endothelial Damage Marker Levels

Jih-Yang Ko1,2,4, Shu-Jui Kuo3, Feng-Sheng Wang1. Ching-Jen Wang1,2, Wen-Yi Chou2, Ka-Kit Siu2 1Department of Medical Research, 2Department of Orthopaedic Surgery, Kaohsiung Chang Gung Memorial Hospital, Taiwan 3Department of Orthopaedic Surgery, China Medical University Hospital, Taichung, Taiwan. 4Department of Orthopaedic Surgery, Xiamen Chang Gung Hospital, Xiamen, Fujian, China

ABSTRACT INTRODUCTION Total knee arthroplasty (TKA) is a well-established modality with a high satisfaction rate for various knee disorders. However, the procedure may be difficult for severe intra-articular or extra-articular deformity of the knee. Besides, this surgical procedure inevitably perturbs the femoral medullary canal, leading to marrow embolization that reportedly increases the risk of myocardial infarction or cardiac stress postoperatively. In addition of improving prosthetic alignment and advantage of feasibility for extra- articular deformity with osteoarthritis, computer-assisted navigation TKAs also contribute to reduced operative blood loss and systemic emboli. These observations imply that navigation TKAs may cause less microvascular damage than conventional TKAs. We postulated that serum levels of CAMs in patients receiving navigation TKAs may be different from those receiving conventional TKAs. MATERIALS AND METHODS We performed a prospective comparative study, enrolling 87 patients from March 2011 to December

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

2011. Fifty-four patients underwent computer navigation TKAs and 33 had conventional TKAs through mid-vastus approach, by two experienced orthopaedic surgeons with the same surgical assistants in the same theater. Levels CAMs, including ICAM-1, VCAM-1, and PECAM-1 in sera and hemovac drainage were measured by ELISA before and 24 hours after surgery. RESULTS There were no significant differences in gender, affected side, age or BMI between the navigation and conventional TKA group. The calculated volume of blood loss in the computer navigation group was 955 (772, 1164) mL, significantly lower (p = 0.001) than the 1265 (963, 1475) mL in the conventional group. Postoperative serum ICAM-1, VCAM-1 and PECAM-1 levels in the navigation group were 35.5% (p < 0.001), 2.0% (p = 0.037) and 49.3% (p < 0.001) lower, than those in the conventional group. The patients in the navigation group had lower ICAM-1 levels in hemovac drainage. DISCUSSION We provided novel evidence that patients had decreased blood loss concomitant with mitigated postoperative elevation of levels of CAMs after navigation TKA. We also reported less elevation of d- Dimer levels after navigation TKA. These are indicative of its less-invasive nature with regard to the integrity of femoral medullary cavity.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Preoperative Templating and Computer-Assisted Total Knee Arthroplasty for Arthritic Valgus Knee

Wen-Yi Chou*, MD, Jih-Yang Ko*†, MD, Ka-Kit Siu*, MD *Department of Orthopedic Surgery,†Department of Medical Research, Kaohsiung Chang Gung Memorial Hospital, Graduate Institute of Clinical Medical Science, Chang Gung University College of Medicine, Kaohsiung, Taiwan

ABSTRACT We reported the functional outcomes, component alignment and optimal thickness of the tibial inserts and joint line changes of 21 arthritic valgus knee deformities using preoperative templating and computer- assisted total knee arthroplasty (TKA). The osseous cut was modified using a novel preoperative templating technique. Soft tissue balance and component implantation were implemented with the aid of a computed tomography-free navigation system. The arthritic valgus knees had clinical, and functional improvement of the knee Society scores and Lysholm scores postoperatively, at an average of 37.8 ± 7.2 months. The mean anatomic axis (15.2° ± 4.5° vs. 6.1° ± 1.4°) and mechanical axis (8.3° ± 5.2° vs. 0.28° ± 1.6°) were also significantly improved postoperatively. The mean thickness of tibial inserts and joint line changes was 10.7 ± 1.46 mm and 0.1 ± 1.4 mm. This computer-assisted technique with preoperative radiographic templating is an alternative strategy to improve TKA results in arthritic valgus knees.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

The Application of Navigation Assisted Total Knee Arthroplasty for Osteoarthritis Knees with Extra-Articular Deformity

Ka-Kit Siu1, Ghe-Wu2, Ching-Jen Wang1, Wen-Yi Chow1, Jih-Yang Koa1,2 1Department of Orthopaedic Surgery, Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Kaohsiung, Taiwan 2Department of Orthopaedic Surgery, Xiamen Chang Gung Hospital, Xiamen, China

ABSTRACT INTRODUCTION For patients with angular extra-articular deformities, retained distal femur internal fixators, narrow femoral canal or long stemmed hip implants, intramedullary guiding may be impossible during conventional TKA. Navigation systems including image-based and image-free system have been used as an adjunct for TKA to improve alignment of the lesion site. These systems can be beneficial for treating knees with severe extra-articular deformities. In this article we demonstrated five successful cases of navigation-assisted TKA for the severe osteoarthritis of the knee with extra-articular deformities or retrained implant, with clinical follow up of three to seven years. MATERIALS AND METHODS There were two males and three females. The average age was 65.6 (range from 49 - 77) at the time of operations. The deformities were located at femur for three patients, tibia for one patient and both femur and tibia for one patient. All the patients underwent navigation assisted TKA, in which both distal femur and proximal tibia cutting and soft tissue balancing were done with the aid of extramedullary CT-free navigation guiding system. RESULTS The average mechanical axis was improved from pre-op varus 14.6 degree (range from varus 3.5 - 20.5 degree) to post-op varus 1.9 ± 0.9 degree (range from varus 0.6 - 3 degree) The average follow up duration was 73.8 months (range from 40-89 months). The average frontal

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016 femoral component (FFC) alignment was valgus 5 ± 1.7 degree and the average lateral femoral component (LFC) alignment was flexion 3 ± 1.5 degree. The average frontal tibia component (FTC) alignment was varus 1 ± 1 degree and the average lateral tibia component (LTC) alignment was posterior 5 ± 2 degrees. The Knee Society Knee Score improved from 46 ± 29 to 94 ± 1 and the function score improved from 40 ± 23 to 79 ± 17. DISCUSSION Our report reveals that navigation system without intramedullary reaming is a feasible and less invasive alternative for complex TKA with severe extra-articular angular deformity or retained implant

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

VII. Invited Session 3 (15:30-17:50, Sept. 5)

Use of Computer Technology in Orthopaedic Surgery and Evaluation of the Surgical Outcome

Shinichi Yoshiya, MD, PhD Department of Orthopaedic Surgery, Hyogo College of Medline, Nishinomiya, Japan

ABSTRACT Primary goal of orthopaedic surgery is restoration of normal anatomy and function of the human locomotive system. In order to achieve this goal, first, normal anatomy and function of the musculoskeletal organs such as joint and spine should be appreciated. Second, abnormalities caused by injuries or diseases need to be evaluated as compared to the normal status. Third, surgical planning is performed to achieve the aim of surgery. In addition, surgical procedure should accurately follow the preoperatively planned steps. Finally, surgical outcome should be evaluated to examine whether the normal anatomy and function can be restored by the surgery. In each process of this sequence, kinematic and image analyses and surgical procedure utilizing computer technology play critical roles to optimize our surgical outcome. In this talk, I would like to present how the orthopaedic surgeons plan and conduct the surgical procedure following the aforementioned steps in knee surgeries such as total knee arthroplasty and ligament reconstruction.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Research and Development in Collaboration with Private Sectors in Biomedical Engineering and Assistive Technology

Toyohiko Hayashi1, PhD, Hiroshi Tanaka2, PhD, Hiroaki Inui2, MD, PhD, Masahiko Komai2, MD, and Katsuya Nobuhara2, MD, PhD 1Graduate School of Science and Technology, Niigata University, Niigata, Japan 2Nobuhara Hospital and Institute of Biomechanics, Hyogo, Japan

ABSTRACT I have been working in interdisciplinary fields of science, such as biomedical and biomechanical engineering, ergonomics and assistive technology. In this presentation, I’d like to demonstrate research and develop projects in collaboration with private sectors. I classified such collaborative approaches into three different categories as follows: 1) university seeds-based type, 2) company needs-based type, and 3) joint development type. As examples of R&D projects belonging to categories 1, 2 and 3, a photoelectric measurement system of 6-DOF jaw movements, TRIMETR JKN/1 (Tokyo Shizaisha Inc.), biomechanical analysis of pitching motion (Nobuhara Hospital) and a web-based management and customizing support system of versatile communication aid for persons with communication disabilities, VCAN/2, under development with Ginzado-Net Inc. were demonstrated, respectively.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

MAKOPlasty Experience of 107 Cases: Robotic Arm Partial Knee Resurfacing

Jiunn-Jer Wu, MD, PhD, Li-Wei Wu, MD, PhD, Ching-Chieh Yang, MD, PhD Chung-Shan Hospital, Taipei, Taiwan

ABSTRACT It is an innovative surgical treatment option for adults living with early to mid-stage Osteoarthritis. It is best suited for patients whose arthritis has not progressed to all three compartments of the knee joint. It is powered by Robotic Arm interactive orthropaedic System; preoperative CT scan of the lesion leg to monitor three dimensional planning of the implants position, intra operative Navigation system to achieve final accurate alignment and precise preparation of the bone, Which allows surgeons to achieve consistent reproducible precision with the newest technique in bone and tissue sparing knee implant surgery. 107 cases was performed from Sep. 2014 to July 2016. The age distribution is from 52 to 94 years old. The Sex ratio is male 34 to female 73. Replaced part is mainly medical compartment 100, lateral compartment 7 cases and only one case with bi-compartment (Medial and PF). Eight cases was replaced on both knee. Sixteen cases has a traditional TKR on the other knee joint. Over all the short term functional results will be shown during the presentation. MAKOplasty is a partial knee replacement with less bone loss and soft tissue injury, preservation of all ligaments and neural elements and accuracy with alignment and position. Early return to normal activity is the end result.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

VIII. Invited Session 4 (09:30-11:30, Sept. 6)

Toward Automated Medical Practice Capsule

Yutaka Hata, PhD Graduate School of Simulation Studies, University of Hyogo, Kobe, Japan

ABSTRACT Currently, a lot of medical and health care studies are done in the world. Especially, many companies focus digital heath care system on cloud computing. On the other side, pure medical issues are also very active in the world. Especially, Japanese government promotes this trend to accelerate economic recovery. In this talk, I will describe automated medical practice capsule in near future. This capsule consists of health care system, automated medical diagnosis system, automated medical treatment system, and automated function recovery system. Almost all medical engineering studies would be an element of it. In them, most difficult task would be a decision making component. This component must make a decision whether the patient must be received medical treatment or not. This decides, what we call, 5W1H, “when” is the user received medical treatment? “what” organ or function is the target? “why” does the user need it? “where” is the best part to do this, “how” does it treat? and “who” corresponds to which is best among robot treatment, drag, or outside of capsule. In it, normality plays primary role to do decide them. I will introduce this normality theory in fuzzy logic framework.

Keywords: medical engineering, automated diagnosis, automated treatment, health, fuzzy logic

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Understanding elderly’s personal differences in participating health-promotion exercise through motion data collection and analysis

Tien-Lung Sun, PhD Department of Industrial Engineering and Management, Yuan-Ze University, Chung-Li, Taiwan

ABSTRACT Research background ： Community-based health-promotion exercises is an important step toward preventive care, which is one of the nine service tasks listed in Taiwan’s new ten-year long-term care 2.0 project. The effectiveness of the health-promotion exercises program is often evaluated by statistical evaluation of clinical measurement. Limitations of clinical measurement is that it only tells the results of the intervention but not the exercise process which leads to the results. Moreover, statistical testing only tells the differences at population level but not at personalized level. The elderly’s personal differences during exercise process, however, is an important factor that affect the final results. Different elderly has different preference and experiences toward a particular exercise program and behave differently based on their personal psychological and physical conditions. If personal differences constantly cause their exercising behavior not achieving the design concept of exercise program, it might lead to negative effects on the quality of the exercise. To make an exercise program intervention successful, it is critical to understand elderly’s personal differences and take proper actions such as communication, care and adjustment. Research goal：The research goal is to collect the elderly’s movement data during health-promotion exercise using wearable sensor and to analyze the data to gain information about the elderly’s personal differences in participating health-promotion exercises.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Method：We have developed a data collection framework which consists of a wearable sensor, a data collection app, cloud storage space, and web-based data analysis and visualization. The inertial sensor in the wearable device is a micro-electro-mechanical system (MEMS) inertial sensors from CavyTech. The sensor is worn by the elderly when conducting health-promotion exercise. The sensor records acceleration and rotation data along X, Y, Z direction with a sampling rate of 30 records per second. The data collection app runs on an Android mobile device. It reads the inertial sensor data through blue tooth and sent the data through Internet to our data storage server installed at Microsoft’s Azure cloud platform. A set of web services were developed and installed at Azure web sites to calculate features from the raw data and to display the raw data and features at web browser. We cooperate with a prototype-testing team from Taiwan’s long-term care 2.0 project to install our data collection framework at 12 elderly health-promotion exercise sites at southern Taiwan. The hardware installation at each site includes 2 wearable sensors, 2 Android mobile devices, a 42-inch TV (to play the health-promotion video), Wi-Fi Internet connection, and other equipment to measure clinical data. We plan to collect motion data during health-promotion exercise to understand the elderly’s personal differences. Data analysis tasks, including feature calculation, feature selection and evaluation, visualization design, statistical testing and machine learning from data like deep learning, and so on are still undergoing.

Keyword：motion data, wearable sensor, personal differences, health promotion exercise

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

The Implementation of Data Visualization Applications in Open-Source Software with Health IT Dashboard Utility

Juei-Chao Chen1, PhD, Chieh-Hsuan Liu1 and Shuen-Ren Cheng2, PhD 1Fu Jen Catholic University 2Cheng Shiu University

ABSTRACT Data visualization is a technique used to explore potential concept in the data by representing it in a chart more suitable for comprehension. However, people who are visually impaired when seeing large quantities of text or numerical characters. But with data visualization science, it aims at helping users to effectively detect the expected, as well as discovering the unexpected to gain insight into the data. Data visualization provides a solution that transforms raw data into visual insight capable of depicting real- time data, future forecasts, trends and growth opportunities. Dashboards are a visual display of the most important information needed to achieve one or more objectives consolidated and arranged on a single screen for real-time data to be monitored at-a-glance. Healthcare dashboard provides healthcare organizations greater insight into their key performance indicators (KPIs) by pulling real-time data from multiple sources into eye-catching, interactive displays. Patented interactive technologies allow key decision makers, board members and trustees to drilldown into charts and graphs and draw real meaning from raw data. In this research, design, implementation, and capabilities of a data visualization system with open source computer language, compatible with multiple database, calculation and graphics library modules, and website-based architecture. The system facilitates the use for developers and users. The System uses Python, MySQL database, Bootstrap, Django framework, and JavaScript. The system ensures that the extension mechanism satisfies the demands of outside developers who wish to incorporate new features and representing the quick display of these data sets for the users with visual charts. The developers can provide answers to data questions, creating a never-ending cycle of long wait times and inflexible results for healthcare organization, including government, hospitals and clinics base on the system architecture. The improved utilization of ‘interactive intelligence’ allows for easy viewing of multiple data points in a single view.

Keyword: data visualization, data dashboard, open source software.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

IX. Invited Session 5 (13:00-15:00, Sept. 6)

Stress Analysis in Various Shoulder Surgeries by Finite Element Analysis

Ching-Chieh Yang, MD, PhD, Juinn-Jer Wu, MD, PhD Department of Orhtopedic Surgery, Chung-Shan Hospital, Taipei, Taiwan

ABSTRACT The finite element method (FEM) is a numerical method to solve the fields problems. FEM has been used in several fields of medical practice and orthopedic practice. It is acknowledged today that FEM will, continuously, contribute to further progress in the design and development of orthopedic implants. In the thesis, we implemented the FEM model to analyze the stress distribution in various shoulder surgeries and investigate the stress distribution in various glenoid designs of reversed shoulder prosthesis (RSP). The shoulder joint is composed of three anatomic components: gleno-humeral (G-H) joint, acromio- clavicular (AC) joint, sterno-clavicular (SC) joint and two main space, namely, the subacromion (SA) space, the scapula-thoracic (ST) space and the coraco-clavicular (C-C) mechanism. Since the author focus on the application of finite element analysis on the realm of shoulder surgery, the anatomic review is emphasized on the detailed structures which are more correlated with surgery of acromio-clavicular joint and design of RSP. With detailed anatomy knowledge, we can construct the finite element model for stress analysis of shoulder surgery. Despite the tremendous progress for shoulder surgery in decades, the surgery for AC joint reconstruction, sequelae of trauma and un-reparable rotator cuff tear remained challenging task for orthopedic surgeon. To our knowledge, this is the first FEA study on AC joint complex. The two dimensional (2D) finite element (FE) models are constructed according to clinically obtained radiographic data and documented parameters of tissues material properties. ANSYS commercial software was adopted to complete the construction process based on finite element method (FEM). The analysis process in

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016 commercial software ANSYS are pre-processor, solver, and post-processor orderly. In pre-processor, choice of proper element type is the first step, followed by entering the material properties of structure. Then, contribute the geometry model and mesh model. In solver, after giving proper boundary conditions, load and solve the problems. As for post-processor, the data and figure could be obtained from the deformation and stress results. In this research, the models are 2D, the element Plane 42 was used. The AC joint Model was then verified by comparing with the published experimental results. The first simulated results showed that coraco-acromial ligament (CAL) plays significant role in biomechanical function of AC joint, and also pointed the numerical support for existence of Salter’s complex. Subsequent studies with this model, we further confirmed that arthroscopic subacromion decompression (ASD) may insult the Salter’s complex and lead to post-operative complication of AC joint. Finally we applied the AC model to simulate distal clavicle resection (DCR). We dictated that the direct method is a better method for DCR and the proper length of DCR is found to be between 0.6 and 0.7 cm. For the study of RSP design, we constructed a model of RSP which is based on the product of Tornier’s company. Simulated results dictated that increased eccentric offset and lateral offset of glenosphere, although being able to reduce scapular notching, may pay the penalty of significant stress concentration over glenoid and its subsequent loosening. Maximum stress occurs at the base of inferior screw elucidates the direct contact failure mode at the middle of inferior screw. Our study provides an alternative tool for the optimal design of glenoid component of RSP. We may further implement the RSP model to find the optimal humeral design and optimal combinations of various designs of RSP in the future. Information about glenohumeral relationships at different angles of abduction is scarce and it remains unclear how angle of arm abduction affects shoulder function. The objective of the study was to clarify relationships between anatomical landmarks of the glenohumeral joint at different angles of abduction.

Keyword：Finite Element, Shoulder Surgery, AC Joint, Reverse Shoulder.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Smartphone-based Statistical Gait Analysis

Chao-Cheng Wu, PhD National Taipei University of Technology, Taipei, Taiwan

ABSTRACT Gait is an important factor for health care and medical diagnosis. Many researches have been focused on this topic, which tried to exploit signals of gaits for many purposes. In neurosurgery, gaits can be regarded as the coordinated action of nerves and muscles. With increase of age, between the strength of muscle and speed of transmission of nerves would be reduced. Additionally, gait could also be an indicator as a nervous system injury or degree of degeneration. Hence, it is critical to build an objective and quantifying method of gait analysis. Traditional gait analysis methods were often confined in the laboratory environment to record the actions of human gait by the sophisticated instruments. These methods are difficult to reflect actions in daily life. We would like to provide a solution by investigating the smartphone-based gait analysis methods. The proposed method exploits signals from the tri-axial accelerometer embedded inside a smartphone as the major sensor. The analysis first recognized seven stages of each gait cycle from signals and then applied Hilbert Transform for frequency-time domain signal analysis. Four major methods to decompose the original gait signals would be investigated, which included Empirical Mode Decomposition (EMD), Multivariate Empirical Mode Decomposition (MEMD), Independent components analysis (ICA), Principal components analysis (PCA). Hilbert Transform was applied later to conduct the frequency-time domain analysis. A few statistical indicators were used for quantitative gait analysis. There are two major contributions. First one is to explore the limitation of smartphone-based gait analysis. Second one aimed to provide important gait information which is difficult to be sensed by human being including medical doctors and a second opinion for medical diagnosis. According to the experimental studies, the proposed method could reveal important gait information even with very low-level smartphone. As a result, the proposed method could also be a long term recovery indicator for doctors and patients.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

The Innovation of Orthopedic Surgery by the Development of In Vivo 3D Kinematic Analysis System

Kazuomi Sugamoto, MD, PhD Orthopaedic Biomaterial Science, Osaka University, Suita City, Japan

ABSTRACT PURPOSE One of the purposes of orthopedic surgeons is a joint recovery of the patients which consists of a pain relief and a normal function of the joints. The evaluation of the function is essential to estimate the clinical results, and it is enabled by the analysis of the joint kinematics. Thousands of cadaveric studies were already reported but those may differ from the in vivo condition for lack of ligament or muscular activity. Plain X- ray only gives us 2D information and it is not sufficient for the 3D analysis. 3D CT or MRI are only the device to visualize the structure under the static conditions. Development of in vivo 3D kinematic analysis system is really needed for the diagnosis of pathological movement or the evaluation of postoperative function. MATERIALS AND METHODS Two systems were already developed in our institute. One is a system using 3D CT or MRI. The targeted joint is placed in serial positions of the motion plane to evaluate 3D kinematics of the motion, and the images are obtained in each position. The data are saved and transmitted to a computer workstation, where image processing such as segmentation and voxel based registration are performed using a software program developed in our laboratory. Segmentation is defined as extracting the contour of the bone required for processing, which is semi-automatically segmented by intensity thresholding techniques. Voxel based registration is an image processing method for matching volume images based on voxel values. Kinematic variables are measured by automatically superimposing segmented 3D MRI or CT of the bones in the neutral position over images for each position using this method of registration. The correlation coefficient is used as a method of measuring similarity of voxel values for registration.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

RESULTS With this method of measuring, a matrix that allowed for the maximal correlation of the two images is calculated. Animations of the joint movement are created from the calculated motions and surface bone models that were reconstructed from 3D-MRI data using the marching cubes algorithm. This method is available for the in vivo 3D kinematic analysis of any joint. The other is a system using a radiographic image intensifier. It can evaluate real-time 3D dynamic motion of the metal implant and it is very available to evaluate the kinematics after the arthroplasty. The 3D pose-estimation technique is built on a 2D/3D registration algorithm, which determines the spatial pose for each component from the implant contours and computer-assisted design (CAD) models of the implant. Sequential fluoroscopic images are taken in the sagittal plane during loaded knee bending motion, walking or stairing. The difference of kinematics in various types of implants were clarified. CONCLUSION More than 3000 data of normal or pathological conditions of all joint were accumulated from all over the world by the collaboration of more than 80 institutes. Pre surgical planning or post-surgical evaluation were dramatically changed. I will show the representative cases, and you can feel how valuable and imperative the image processing technique or 3D kinematic analysis is for the orthopedic clinical fields. We also launched the iPad or Google play application for the education of 4D human anatomy.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

X. General Session 2 (15:20-16:40, Sept. 6)

Interconnection between Stretchable Wiring and Integrated Circuit Using Magnetic Attractive Force

Motoki Tsukiyama Dept. of Electrical Engineering and Computer Science, University of Hyogo, Himeji, Japan

ABSTRACT Along with developing the preventive medicine, various devices for sensing biomedical information have been researched. By pasting the sensor onto the human skin directly, the sensor obtains accurate biomedical data. The wearable devices should be more flexible than conventional devices. The electric connection between flexible components and rigid integrated circuit (IC) chips is easy to break by stress concentration. Therefore, the durable connection between flexible and rigid materials is required. In this study, we demonstrate a novel interconnection that is the buried IC and stretchable wiring in silicone rubber. The structure of this interconnection is composed by magnetic powder wiring and the magnetic electrode. The magnetic attractive force generated from this interconnection enables this interconnection to have self-healing property. This interconnection is able to apply various sensing systems.

Keywords: wearable device, stretchable wiring, silicone rubber, magnetic attractive force, self-healing property

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Wearable Pulse Sensing System with P(VDF/TrFE) Piezoelectric Film for Health Management

Ryosuke Fujita Dept. of Electrical Engineering and Computer Science, University of Hyogo, Himeji, Japan

ABSTRACT Wearable health monitoring systems have been attracting attention to manage human health conditions. Vital signs, e.g. blood pressure, heart rate, breathing and body temperature are important for health management. In order to monitor real-time vital signs, wearable devices are required flexibility to paste directly on human skin. Therefore, we propose a flexible and wearable pulse sensing system by using polyvinylidene fluoride-trifluoroethylene(P(VDF/TrFE)) piezoelectric film sensor. We fabricated the piezoelectric sensor with P(VDF/TrFE) film sandwiched between two gold electrodes that are covered the whole with a silicone rubber. The system is contained with a PSoC microcontroller that has analog and digital signal processing block and Bluetooth Low Energy block. In the conference, the fabrication step and preliminary evaluation results and possibility to measure the heart rate, blood pressure and breathing will be presented.

Keywords: piezoelectric film, wearable pulse sensing, flexible electronics, health monitoring

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Biomechanical Approach to the Shoulder Complex System

Hiroshi Tanaka1, PhD, Toyohiko Hayashi2, PhD, Hiroaki Inui1, MD, PhD, Shyoji Kobashi3, PhD, Shinich Yoshiya4, MD, PhD, Koji Shibano5, MD, PhD, and Katsuya Nobuhara1, MD, PhD 1Nobuhara Hospital and Institute of Biomechanics, Tatsuno, Japan 2Graduate School of Science and Technology, Niigata University, Niigata, Japan 3Graduate School of Engineering, University of Hyogo, Himeji, Japan 4Department of Orthopaedic Surgery, Hyogo College of Medline, Nishinomiya, Japan 5Orthopaedic Biomaterial Science, Osaka University, Suita, Japan

ABSTRACT The shoulder joint is an elusive system. Quantitative assessment of this complex system is significantly challenging for understanding musculoskeletal injury, disease and treatment. Biomechanics would provide a useful clue to quantitative analysis of how the shoulder complex moves and functions. Up-to-date biomechanical studies provide three-dimensional kinematics and kinetics assessments to describe shoulder complex motion frequently-utilizing fresh frozen cadaver, computed tomography (CT), magnetic resonance imaging (MRI), x-ray fluoroscopy, electromagnetic tracking device or three- dimensional marker-based automatic digitizing motion capture system or video system. These instruments and devices have features and shortcomings, such as in vivo or in vitro, invasive or non-invasive, dynamic or static/semi-static, etc. Additionally, in order to determine joint function at a level relevant to specific joint structures, high accuracy (±1 mm and degree or better) is required. Simulation of the shoulder complex movable in the virtual world is also a valuable computational method. When simulation is combined with biomechanical analysis and anatomical modeling to investigate joint system and soft tissue mechanics at the structural level, it is possible to visualize the realistic results under static, dynamic and living environments in animated forms. We believe that in vivo and in vitro functional joint assessment with biomechanical and simulating technique provide a powerful tool for establishing kinematic and kinetic basis for the development and evaluation of new treatments that may be more effective for optimizing functional outcomes and preventing various shoulder injuries. This presentation shows specific three-dimensional shoulder motion pattern in vivo and in vitro.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Glenohumeral Relationships at Different Angles of Shoulder Abduction

Hiroaki Inui, MD, PhD, Hiroshi Tanaka, PhD and Katsuya Nobuhara, MD, PhD Nobuhara Hospital and Institute of Biomechanics, Tatsuno, Japan

ABSTRACT PURPOSE: Information about glenohumeral relationships at different angles of abduction is scarce and it remains unclear how angle of arm abduction affects shoulder function. The objective of the study was to clarify relationships between anatomical landmarks of the glenohumeral joint at different angles of abduction. METHODS: Fifteen volunteers (10 men, 5 women; mean age, 31 years) were enrolled in this study. Images of externally and internally rotated positions at 45°, 90°, and 135° of abduction in the scapular plane were taken using an open MRI system, and landmarks including glenoidal long axis, center of the head, bicipital groove, and humeral shaft axis were determined. Using the line set on the head surface by the plane parallel to the humeral axis, involving the head center and the bicipital groove with its parallel and perpendicular lines, glenoid location and rotational relationships were investigated in each position. RESULTS: Angle values of axial rotation averaged 48° at 45° of abduction, 71° at 90° of abduction, and 40° at 135° of abduction. Humeral heads remained centered and there was no difference in location of the head centers on the glenoid between the six positions. The trajectories of glenoid center extended primarily over anterior portion of the humeral head at 45° of abduction and extended over posterior portion at 90° of abduction, while those at 135° of abduction were localized on a small upper portion of the head. CONCLUSIONS: The glenohumeral relationships demonstrated angle of abduction might influence shoulder function through its effects on the portion of the humeral surface in contact with the glenoid during rotation and the resultant changes in the glenohumeral relationships.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

XI. General Session 3 (16:50-17:50, Sept. 6)

Energy Harvesting in IoT Sensors

Takayuki Fujita, PhD Graduate School of Engineering, University of Hyogo, Himeji, Japan

ABSTRACT Recently, much attention has been placed on the “IoT (Internet of Things)” and “Trillion Sensor Universe” markets and technologies. IoT refers not only to computers and mobile phones connection through the Internet, but to the wireless interconnection of all of the huge number of “things” and devices especially the wireless sensor node (WSN). Additionally, Trillion Sensor Universe refers expansion of sensor numbers to trillions. With the rapid development of the ultra-low power WSN, the autonomous energy sources e.g. maintenance free battery and energy harvester are attracting as a most important issue. In this presentation, the brief introduction of the IoT with related technical circumference and MEMS (Micro ElectroMechanical Systems) based energy harvesters for sensors will be presented. The vibration energy harvester is one of the autonomous power source that obtain electrical energy from the vibration on human or machine motions etc.

Keywords: IoT, Internet of Things, sensors, energy harvester, autonomous power source

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

Post-Operative Implanted Knee Kinematics Prediction for Personalized Medicine

Syoji Kobashi1, Manabu Nii1, Belayat Hossain1, Syunichiro Kambara2, Takatoshi Morooka2, Makiko Okuno2, Shinichi Yoshiya2 1Graduate School of Engineering, University of Hyogo, Himeji, Japan 2Hyogo College of Medicine, Kobe, Japan

ABSTRACT Personalized medicine has attracted considerable recent attention across many areas of clinical practice because it will increase the outcome of treatment and may decrease the medical cost. This study aims to proceed personalized medicine in total knee arthroplasty (TKA) surgery, which replaces the damaged knee joint with artificial ones called implant. Especially, we focus on prediction of post-operative implanted knee function in TKA surgery for the sake of assisting orthopedicians to select an appropriate TKA implant type which maximizes the outcome of treatment for individual patient. The proposed method is based on big data and machine learning algorithms. It extracts features using principal component analysis, and constructs a mapping function from pre-operative feature space to post-operative feature space. The method was validated using by applying to prediction of post-operative anterior-posterior translation in 52 TKA operated knees. Leave-one-out cross validation test revealed the prediction performances with a mean correlation coefficients of 0.79 and a mean root-mean-squared-error of 3.44 mm.

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016

On Personal Identification Using Clinical Data Obtained by Optical Coherence Tomography

N. Kamiura1, S. Kobashi1, M. Nii1, T. Yumoto1, T. Yamauchi2, H. Tabuchi2 1Graduate School of Engineering, University of Hyogo, Himeji, Japan 2 Department of Ophthalmology, Tsukazaki Hospital, Himeji, Japan

ABSTRACT The personal identification is presented, using the data associated with corneal thickness obtained by Optical Coherence Tomography (OCT). This OCT provides matrices, each of which consists of thirty-two rows and two-hundred and fifty-six columns, to users as output sequences for test items. The proposed method generates two thirty-two-dimensional vectors (or a sixty-four-dimensional vector) from the matrix associated with a test item for a test subject, and adds it to a set of registered data. When identifying some -sized matrix by OCT and generates the collation data as well as when registering the corresponding subject. This OCT does not always measure the test item value at the identical point on the cornea. In other words, the point at which the item value is measured when generating the registered data tends to differ from the point at which the item value is measured when generating the collation data, even if these item values are specified by the same row and the same column in the two matrices. To cope with the issue related to the gap occurring on the cornea when generating registration data and collation data, the proposed method focuses on test items of which values are not only enough different among subjects but also have comparatively low dependency on measuring points for the identical subject. It employs Axial Keratometric, Axial Posterior, and Pachymetry as candidate items fulfilling the above restriction, and examines the identification rate achieved when registration and collation data are generated f -sized matrix for each of the three items. The proposed method picks up the maximum value and the minimum value on each row in the matrix corresponding to one of the test items, and generates the data with the maximum value and/or the minimum value for elements. The three types of the distance are defined, and the proposed method calculates the distance between each of the members in the registered data set and the given collation data. It judges the test subject corresponding to the given collation data to be that of the registered data with the shortest

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016 distance to the given data. Experimental results establish that the proposed method achieves 100 percent as the identification rate, provided that registered and collation data are generated from the matrix associated with Pachymetry and that the number of subjects is at most 45.

Keywords: Personal Identification; Optical Coherence Tomography; Cornea; Pachymetry

The 6th International Symposium in Computational Medical and Health Technology Kaohsiung, Taiwan, Sept. 5-7, 2016 XII. Contact

JAPAN TEAM: Naotake KAMIURA, 上浦尚武 University of Hyogo, Himeji, Japan Email: [email protected]

Hiroshi TANAKA, 田中洋 Nobuhara Hospital and Institute of Biomechanics Email: [email protected]

Syoji KOBASHI, 小橋昌司 University of Hyogo, Himeji, Japan Email: [email protected]

TAIWAN TEAM: Chao-Cheng Wu, 吳昭正 Dept. of Electrical Engineering, Taipei University of Technology Email: [email protected]

Juei-Chao Chen, 陳瑞照 Dept. of Statistics and Information Science, Fu Jen Catholic University Email: [email protected]

Lung-Yu Li, 李龍雨 Cheng Shiu University, Kaohsiung, Taiwan Email: [email protected]

ＵｎｉｖｅｒｓｉｔｙｏｆＨＹＯＧＯ＆ＣｈｅｎｇＳｈｉｕＵｎｉｖｅｒｓｉｔｙ