DOCSLIB.ORG

Dominic Jack Thesis

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Dominic Jack Thesis Deep Learning Approaches for 3D Inference from Monocular Vision SUBMITTED IN FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY Dominic Jack Bachelor of Applied Science (Hons) School of Electrical Engineering and Robotics Science and Engineering Faculty Queensland University of Technology 2020 Copyright in Relation to This Thesis c Copyright 2020 by Dominic Jack Bachelor of Applied Science (Hons). All rights reserved. Statement of Original Authorship The work contained in this thesis has not been previously submitted to meet requirements for an award at this or any other higher education institution. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made. Signature: QUT Verified Signature Date: 3 September 2020 i ii Abstract Deep learning has contributed significant advances to computer vision in the last decade. This thesis looks at two problems involving 3D inference from 2D inputs: human pose estimation, and single-view object reconstruction. Each of our methods considers a different type of 3D representation that seeks to take advantage of the representation’s strengths, including keypoints, occupancy grids, deformable meshes and point clouds. We additionally investigate methods for learning from unstructured 3D data directly, including point clouds and event streams. In particular, we focus on methods targeted towards applications on moderately-sized mobile robotics platforms with modest computational power on board. We prioritize methods that operate in real-time with relatively low memory footprint and power usage compared to those tuned purely for accuracy-like performance metrics. Our first contribution looks at 2D-to-3D human pose keypoint lifting, i.e. how to infer a 3D human pose from 2D keypoints. We use a generative adversarial network to learn a latent space corresponding to feasible 3D poses, and optimize this latent space at inference time to find an element corresponding to the 3D pose which is most consistent with the 2D observation using a known camera model. This results in competitive accuracies using a very small generator model. Our second contribution looks at single-view object reconstruction using deformable mesh models. We learn to simultaneously choose a template mesh from a small number of candidates and infer a continuous deformation to apply to that mesh based on an input image. We tackle both problems of human pose estimation and single-view object reconstruction in our third contribution. Through a reformulation of the model presented in our first contribution, we combine multiple separate optimization steps into a single multi-level optimization problem that takes into account the feasibility of the 3D representation and its consistency with observed 2D features. We show that approximate solutions to the inner optimization process can be expressed as a learnable layer and propose problem-specific networks which we call Inverse Graphics Energy Networks (IGE-Nets). For human pose estimation, we achieve comparable results to benchmark deep learning models with a fraction of iii the number of operations and memory footprint, while our voxel-based object reconstruction model achieves state-of-the-art results at high resolution on a standard desktop GPU. Our final contribution was initially intended to extend our IGE-Net architecture to accommodate point clouds. However, a search of the literature found no simple network architectures which were both hierarchical in cloud density and continuous in coordinates – both necessary conditions for efficient IGE- Nets. As such, we present various approaches that improve performance of existing point cloud methods, and present a modification which is not only hierarchical and continuous, but also runs significantly faster and requires significantly less memory than existing methods. We further extend this work for use with event camera streams, producing networks that take advantage of the asynchronous nature of the input format and achieve state-of-the-art results on multiple classification benchmarks. iv Acknowledgments To my supervisors, thank you for your guidance and the freedom to chase my ideas – even when you knew they would not lead anywhere; to other students and academics in the lab, thank you for the advice, the water-cooler chats, and listening to my rants of frustrations; to my parents, thank you for your unquestioning love and support whether I’m starting a circus school or finishing a PhD; and finally to my friends, who kept me sane, accepted when I was distant, and welcomed me back after each of my deadline-induced disappearances. This project has pushed me to my limit, and I could not have got through without each and every one of you. I would also like to thank the Australian government for the Research Training Program (RTP) of which I was a recipient, along with QUT’s Deputy Vice-Chancellor’s Initiative Scholarship. v vi Table of Contents Abstract iii Acknowledgments v Acronyms xi List of Figures xiii List of Publications 1 1 Introduction 3 1.1 Computer Vision in Robotics . .4 1.2 Artificial Intelligence and the Deep Learning Era . .5 1.3 Problem Descriptions . .6 1.3.1 Human Pose Estimation . .6 1.3.2 Single View Object Reconstruction . .8 1.3.3 Point Cloud Classification . 10 1.3.4 Event Stream Classification . 10 1.4 Research Questions . 11 1.5 Thesis Outline . 11 2 Literature Review 13 2.1 General Deep Learning . 13 2.1.1 The Exploding/Vanishing Gradient Problem . 13 2.1.2 Optimizers . 14 vii 2.1.3 Initialization . 15 2.1.4 Regularization . 15 2.1.5 Implementation and Training . 16 2.2 Generative Adversarial Networks . 18 2.3 Learned Energy Networks . 19 2.4 Deep Learning in Computer Vision . 19 2.4.1 Object detection . 20 2.4.2 Semantic Segmentation . 20 2.4.3 Instance Segmentation . 21 2.4.4 Human Pose Estimation . 21 2.4.5 Single View Object Reconstruction . 22 2.4.6 Point Cloud Networks . 23 2.4.7 Event Stream Networks . 24 3 Adversarially Parameterized Optimization for 3D Human Pose Estimation 27 4 Learning Free-Form Deformations for 3D Object Reconstruction 41 5 IGE-Net: Inverse Graphics Energy Networks for Human Pose Estimation and Single-View Reconstruction 59 6 Sparse Convolutions on Continuous Domains for Point Cloud and Event Stream Networks 73 7 Conclusion 99 7.1 Contributions . 99 7.2 Future Work . 100 7.2.1 Sparse Point Cloud Convolutions . 100 7.2.2 Sparse Event Stream Convolutions . 100 7.2.3 IGE-Nets . 101 7.2.4 Isosurface Extraction Models . 101 7.2.5 Universal Latent Shape Representation . 102 viii References 118 ix x Acronyms AI Artificial Intelligence DL Deep Learning CNN Convolutional Neural Network CPU Central Processing Unit FFD Free Form Deformation GAN Generative Adversarial Network GPU Graphics Processing Unit IGE-Net Inverse Graphics Energy Network ML Machine Learning NLP Natural Language Processing QUT Queensland University of Technology ReLU Rectified Linear Units R-CNN Region-based CNN SGD Stochastic Gradient Descent TPU Tensor Processing Unit xi xii List of Figures 1.1 Left-to-right: fleets of driverless cars are already being built [1]; doctor-assisted medical diagnosis improves accuracy and reduces doctor stress and fatigue [2]; a robotic “dog” is in use at San Francisco airport, taking and collating construction photographs [3]; Samsung’s SRG-A1 sentry gun on the South Korean side of the demilitarized zone [4]. .4 1.2 Left-to-right: image classification, object detection and localization, semantic segmenta- tion, and instance segmentation are common tasks in 2D computer vision. .4 1.3 Human pose representations. Left-to-right: 2D heatmap and keypoint detections [5] and 3D keypoint detections [6]. .7 1.4 3D object representations. Left-to-right: triangular mesh, point cloud, voxel occupancy grid, level sets slice. .9 1.5 Output from an event camera can be thought of as a point-cloud in x; y; t space. 10 xiii xiv List of Publications Following is the list of peer-reviewed publications that form a part of this thesis. Dominic Jack, Frederic Maire, Anders Eriksson, and Sareh Shirazi. “Adversarially Parameterized • Optimization for 3D Human Pose Estimation” in International Conference on 3D Vision (3DV). Qingdao, China, 2017. Dominic Jack, Jhony K. Pontes, Sridha Sridharan, Clinton Fookes, Sareh Shirazi, Frederic Maire, • Anders Eriksson. “Learning Free-Form Deformations for 3D Object Reconstruction,” in the Asian Conference on Computer Vision (ACCV). Perth, Australia, 2018. Dominic Jack, Frederic Maire, Sareh Shirazi, and Anders Eriksson. “IGE-Net: Inverse Graphics • Energy Networks for Human Pose Estimation and Single-View Reconstruction,” in the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Long Beach, USA, 2019. Dominic Jack, Frederic Maire, Simon Denman, and Anders Eriksson. “Sparse Convolutions on • Continuous Domains for Point Cloud and Event Stream Networks,” submitted to the European Conference on Computer Vision (ECCV). Glasgow, UK, 2020. The following publication is related to this thesis but not included. Mateusz Michalkiewicz, Jhony K. Pontes, Dominic Jack, Mahsa Baktashmotlagh, Anders Eriks- • son. “Implicit Surface Representations As Layers in Neural Networks,” in International Confer- ence on Computer Vision (ICCV). Seoul, Korea, 2019. 1 2 Chapter 1 Introduction “As I dug into research on Artificial Intelligence, I could not believe what I was reading. It hit me pretty quickly that what’s happening in the world of AI is not just an important topic, but by far THE most important topic for our future.” — Tim Urban, Wait but Why [7] Driverless cars. Aerial drone-based parcel delivery. Robots doing house work. Military sentry guns automatically selecting and eliminating targets. These aren’t concepts discussed by futurists predicted for the far-off future, or even experimental proof-of-concept implementations in highly controlled labs. All are in operation today across the globe, and progress in the area is only speeding up.
Load more

Recommended publications
  • Image-Based 3D Reconstruction: Neural Networks Vs. Multiview Geometry
    Image-based 3D Reconstruction: Neural Networks vs. Multiview Geometry Julius Schoning¨ and Gunther Heidemann Institute of Cognitive Science, Osnabruck¨ University, Osnabruck,¨ Germany Email: fjuschoening,[email protected] Abstract—Methods using multiple view geometry (MVG), like algorithms, guarantee linear processing time, even in cases Structure from Motion (SfM), are still the dominant approaches where the number and resolution of the input images make for image-based 3D reconstruction. These reconstruction methods MVG-based approaches infeasible. have become quite robust and accurate. However, how robust and accurate can artificial neural networks (ANNs) reconstruct For the research if the underlying mathematical principle a priori unknown 3D objects? Exceed the restriction of object of MVG can be learned by ANNs, datasets like ShapeNet categories this paper evaluates ANNs for reconstructing arbitrary [9] and ModelNet40 [10] cannot be used hence they en- 3D objects. With the use of a synthetic scalable cube dataset for code object categories such as planes, chairs, tables. The training, testing and validating ANNs, it is shown that ANNs are needed dataset must not have shape priors of object cat- capable of learning mathematical principles of 3D reconstruction. As inspiration for the design of the different ANNs architectures, egories, and also, it must be scalable in its complexity the global, hierarchical, and incremental key-point matching for providing a large body of samples with ground truth strategies of SfM approaches were taken into account. Based data, ensuring the learning of even deep ANN. For this on these benchmarks and a review of the used dataset, it is reason, we are using the synthetic scalable cube dataset [11], shown that voxel-based 3D reconstruction cannot be scaled.
    [Show full text]
  • Amodal 3D Reconstruction for Robotic Manipulation Via Stability and Connectivity
    Amodal 3D Reconstruction for Robotic Manipulation via Stability and Connectivity William Agnew, Christopher Xie, Aaron Walsman, Octavian Murad, Caelen Wang, Pedro Domingos, Siddhartha Srinivasa University of Washington fwagnew3, chrisxie, awalsman, ovmurad, wangc21, pedrod, [email protected] Abstract: Learning-based 3D object reconstruction enables single- or few-shot estimation of 3D object models. For robotics, this holds the potential to allow model-based methods to rapidly adapt to novel objects and scenes. Existing 3D re- construction techniques optimize for visual reconstruction fidelity, typically mea- sured by chamfer distance or voxel IOU. We find that when applied to realis- tic, cluttered robotics environments, these systems produce reconstructions with low physical realism, resulting in poor task performance when used for model- based control. We propose ARM, an amodal 3D reconstruction system that in- troduces (1) a stability prior over object shapes, (2) a connectivity prior, and (3) a multi-channel input representation that allows for reasoning over relationships between groups of objects. By using these priors over the physical properties of objects, our system improves reconstruction quality not just by standard vi- sual metrics, but also performance of model-based control on a variety of robotics manipulation tasks in challenging, cluttered environments. Code is available at github.com/wagnew3/ARM. Keywords: 3D Reconstruction, 3D Vision, Model-Based 1 Introduction Manipulating previously unseen objects is a critical functionality for robots to ubiquitously function in unstructured environments. One solution to this problem is to use methods that do not rely on explicit 3D object models, such as model-free reinforcement learning [1,2]. However, quickly generalizing learned policies across wide ranges of tasks and objects remains an open problem.
    [Show full text]
  • Stereoscopic Vision System for Reconstruction of 3D Objects
    International Journal of Applied Engineering Research ISSN 0973-4562 Volume 13, Number 18 (2018) pp. 13762-13766 © Research India Publications. http://www.ripublication.com Stereoscopic Vision System for reconstruction of 3D objects Robinson Jimenez-Moreno Professor, Department of Mechatronics Engineering, Nueva Granada Military University, Bogotá, Colombia. Javier O. Pinzón-Arenas Research Assistant, Department of Mechatronics Engineering, Nueva Granada Military University, Bogotá, Colombia. César G. Pachón-Suescún Research Assistant, Department of Mechatronics Engineering, Nueva Granada Military University, Bogotá, Colombia. Abstract applications of mobile robotics, for autonomous anthropomorphic agents [9] or not, require reducing costs and The present work details the implementation of a stereoscopic using free software for their development. Which, by means of 3D vision system, by means of two digital cameras of similar laser systems or mono-camera, does not allow to obtain a characteristics. This system is based on the calibration of the reduction in price or analysis of adequate depth. two cameras to find the intrinsic and extrinsic parameters of the same in order to use projective geometry to find the disparity The system proposed in this article presents a system of between the images taken by each camera with respect to the reconstruction with stereoscopic pair, i.e. two cameras that will same scene. From the disparity, it is obtained the depth map capture the scene from their own perspective allowing, by that will allow to find the 3D points that are part of the means of projective geometry, to establish point reconstruction of the scene and to recognize an object clearly correspondences in the scene to find the calibration matrices of in it.
    [Show full text]
  • Configurable 3D Scene Synthesis and 2D Image Rendering with Per-Pixel Ground Truth Using Stochastic Grammars
    International Journal of Computer Vision https://doi.org/10.1007/s11263-018-1103-5 Configurable 3D Scene Synthesis and 2D Image Rendering with Per-pixel Ground Truth Using Stochastic Grammars Chenfanfu Jiang1 · Siyuan Qi2 · Yixin Zhu2 · Siyuan Huang2 · Jenny Lin2 · Lap-Fai Yu3 · Demetri Terzopoulos4 · Song-Chun Zhu2 Received: 30 July 2017 / Accepted: 20 June 2018 © Springer Science+Business Media, LLC, part of Springer Nature 2018 Abstract We propose a systematic learning-based approach to the generation of massive quantities of synthetic 3D scenes and arbitrary numbers of photorealistic 2D images thereof, with associated ground truth information, for the purposes of training, bench- marking, and diagnosing learning-based computer vision and robotics algorithms. In particular, we devise a learning-based pipeline of algorithms capable of automatically generating and rendering a potentially infinite variety of indoor scenes by using a stochastic grammar, represented as an attributed Spatial And-Or Graph, in conjunction with state-of-the-art physics- based rendering. Our pipeline is capable of synthesizing scene layouts with high diversity, and it is configurable inasmuch as it enables the precise customization and control of important attributes of the generated scenes. It renders photorealistic RGB images of the generated scenes while automatically synthesizing detailed, per-pixel ground truth data, including visible surface depth and normal, object identity, and material information (detailed to object parts), as well as environments (e.g., illuminations and camera viewpoints). We demonstrate the value of our synthesized dataset, by improving performance in certain machine-learning-based scene understanding tasks—depth and surface normal prediction, semantic segmentation, reconstruction, etc.—and by providing benchmarks for and diagnostics of trained models by modifying object attributes and scene properties in a controllable manner.
    [Show full text]
  • 3D Scene Reconstruction from Multiple Uncalibrated Views
    3D Scene Reconstruction from Multiple Uncalibrated Views Li Tao Xuerong Xiao [email protected] [email protected] Abstract aerial photo filming. The 3D scene reconstruction applications such as Google Earth allow people to In this project, we focus on the problem of 3D take flight over entire metropolitan areas in a vir- scene reconstruction from multiple uncalibrated tually real 3D world, explore 3D tours of build- views. We have studied different 3D scene recon- ings, cities and famous landmarks, as well as take struction methods, including Structure from Mo- a virtual walk around natural and cultural land- tion (SFM) and volumetric stereo (space carv- marks without having to be physically there. A ing and voxel coloring). Here we report the re- computer vision based reconstruction method also sults of applying these methods to different scenes, allows the use of rich image resources from the in- ranging from simple geometric structures to com- ternet. plicated buildings, and will compare the perfor- In this project, we have studied different 3D mances of different methods. scene reconstruction methods, including Struc- ture from Motion (SFM) method and volumetric 1. Introduction stereo (space carving and voxel coloring). Here we report the results of applying these methods to 3D reconstruction from multiple images is the different scenes, ranging from simple geometric creation of three-dimensional models from a set of structures to complicated buildings, and will com- images. It is the reverse process of obtaining 2D pare the performances of different methods. images from 3D scenes. In recent decades, there is an important demand for 3D content for com- 2.
    [Show full text]
  • 3D Shape Reconstruction from Vision and Touch
    3D Shape Reconstruction from Vision and Touch Edward J. Smith1;2∗ Roberto Calandra1 Adriana Romero1;2 Georgia Gkioxari1 David Meger2 Jitendra Malik1;3 Michal Drozdzal1 1 Facebook AI Research 2 McGill University 3 University of California, Berkeley Abstract When a toddler is presented a new toy, their instinctual behaviour is to pick it up and inspect it with their hand and eyes in tandem, clearly searching over its surface to properly understand what they are playing with. At any instance here, touch provides high fidelity localized information while vision provides complementary global context. However, in 3D shape reconstruction, the complementary fusion of visual and haptic modalities remains largely unexplored. In this paper, we study this problem and present an effective chart-based approach to multi-modal shape understanding which encourages a similar fusion vision and touch information. To do so, we introduce a dataset of simulated touch and vision signals from the interaction between a robotic hand and a large array of 3D objects. Our results show that (1) leveraging both vision and touch signals consistently improves single- modality baselines; (2) our approach outperforms alternative modality fusion methods and strongly benefits from the proposed chart-based structure; (3) the reconstruction quality increases with the number of grasps provided; and (4) the touch information not only enhances the reconstruction at the touch site but also extrapolates to its local neighborhood. 1 Introduction From an early age children clearly and often loudly demonstrate that they need to both look and touch any new object that has peaked their interest. The instinctual behavior of inspecting with both their eyes and hands in tandem demonstrates the importance of fusing vision and touch information for 3D object understanding.
    [Show full text]
  • 3D Reconstruction Is Not Just a Low-Level Task: Retrospect and Survey
    3D reconstruction is not just a low-level task: retrospect and survey Jianxiong Xiao Massachusetts Institute of Technology [email protected] Abstract 3D reconstruction is in obtaining more accurate depth maps [44, 45, 55] or 3D point clouds [47, 48, 58, 50]. We now Although an image is a 2D array, we live in a 3D world. have reliable techniques [47, 48] for accurately computing The desire to recover the 3D structure of the world from 2D a partial 3D model of an environment from thousands of images is the key that distinguished computer vision from partially overlapping photographs (using keypoint match- the already existing field of image processing 50 years ago. ing and structure from motion). Given a large enough set of For the past two decades, the dominant research focus for views of a particular object, we can create accurate dense 3D reconstruction is in obtaining more accurate depth maps 3D surface models (using stereo matching and surface fit- or 3D point clouds. However, even when a robot has a depth ting [44, 45, 55, 58, 50, 59]). In particular, using Microsoft map, it still cannot manipulate an object, because there is Kinect (also Primesense and Asus Xtion), a reliable depth no high-level representation of the 3D world. Essentially, map can be obtained straightly out of box. 3D reconstruction is not just a low-level task. Obtaining However, despite all of these advances, the dream of hav- a depth map to capture a distance at each pixel is analo- ing a computer interpret an image at the same level as a two- gous to inventing a digital camera to capture the color value year old (for example, counting all of the objects in a pic- at each pixel.
    [Show full text]
  • Automatic Reconstruction of Textured 3D Models of Textured 3Dmodels Automatic Reconstruction Dipl.-Ing
    Dipl.-Ing. Benjamin Pitzer Automatic Reconstruction of Textured 3D Models Automatic Reconstruction of Textured 3D Models Automatic Reconstruction of Textured Benjamin Pitzer 020 Benjamin Pitzer Automatic Reconstruction of Textured 3D Models Schriftenreihe Institut für Mess- und Regelungstechnik, Karlsruher Institut für Technologie (KIT) Band 020 Eine Übersicht über alle bisher in dieser Schriftenreihe erschienenen Bände finden Sie am Ende des Buchs. Automatic Reconstruction of Textured 3D Models by Benjamin Pitzer Dissertation, Karlsruher Institut für Technologie (KIT) Fakultät für Maschinenbau Tag der mündlichen Prüfung: 22. Februar 2011 Referenten: Prof. Dr.-Ing. C. Stiller, Adj. Prof. Dr.-Ing. M. Brünig Impressum Karlsruher Institut für Technologie (KIT) KIT Scientific Publishing Straße am Forum 2 D-76131 Karlsruhe KIT Scientific Publishing is a registered trademark of Karlsruhe Institute of Technology. Reprint using the book cover is not allowed. www.ksp.kit.edu This document – excluding the cover – is licensed under the Creative Commons Attribution-Share Alike 3.0 DE License (CC BY-SA 3.0 DE): http://creativecommons.org/licenses/by-sa/3.0/de/ The cover page is licensed under the Creative Commons Attribution-No Derivatives 3.0 DE License (CC BY-ND 3.0 DE): http://creativecommons.org/licenses/by-nd/3.0/de/ Print on Demand 2014 ISSN 1613-4214 ISBN 978-3-86644-805-6 DOI: 10.5445/KSP/1000025619 Automatic Reconstruction of Textured 3D Models Zur Erlangung des akademischen Grades eines Doktors der Ingenieurwissenschaften von der Fakultät für Maschinenbau der Universität Karlsruhe (TH) genehmigte Dissertation von DIPL.-ING.BENJAMIN PITZER aus Menlo Park, CA Hauptreferent: Prof. Dr.-Ing. C. Stiller Korreferent: Adj.
    [Show full text]
  • Image-Based Synthesis and Re-Synthesis of Viewpoints Guided by 3D Models
    Image-based Synthesis and Re-Synthesis of Viewpoints Guided by 3D Models Konstantinos Rematas1, Tobias Ritschel2,3, Mario Fritz2, and Tinne Tuytelaars1 1KU Leuven, iMinds 2Max Planck Institute for Informatics 2Saarland University Abstract tinely by artists to convey presence, most well-known as the Ken Burn’s effect [15]. We propose a technique to use the structural informa- In contrast, the predominant paradigm in computer vi- tion extracted from a set of 3D models of an object class to sion is to present all possible viewpoints in order to arrive improve novel-view synthesis for images showing unknown at a model that is robust to out-of-plane rotations. The most instances of this class. These novel views can be used to prominent detection models lack the domain knowledge “amplify” training image collections that typically contain that would give them an understanding of how generaliza- only a low number of views or lack certain classes of views tion across viewpoints can be achieved from a single-view entirely (e. g. top views). example. A dense sampling across viewpoints and intra- We extract the correlation of position, normal, re- category variation is tedious to achieve. Recent analysis of flectance and appearance from computer-generated images such detectors has pointed out that rare cases in viewpoint is of a few exemplars and use this information to infer new indeed one of the frontiers on which there is still significant appearance for new instances. We show that our approach room to push the state-of-the-art in object detection [14]. can improve performance of state-of-the-art detectors using In this work, we show how to improve novel view syn- real-world training data.
    [Show full text]
  • 3D Reconstruction and Recognition Acknowledgement
    EE290T : 3D Reconstruction and Recognition Acknowledgement Courtesy of Prof. Silvio Savarese. Introduction “There was a table set out under a tree in front of the house, and the March Hare and the Hatter were having tea at it.” “The table was a large one, but the three were all crowded together at one corner of it …” From “A M a d Tea-Party” Alice's Adventures in W onderla n by Lewis Ca rroll “There was a table set out under a tree in front of the house, and the March Hare and the Hatter were having tea at it.” “The table was a large one, but the three were all crowded together at one corner of it …” From “A M a d Tea-Party” Alice's Adventures in W onderla n by Lewis Ca rroll Illustra tion by Arthur Ra ck ha m Computer vision Image/ video Object 1 Object N - semantic -semantic … Computer vision Image/ video Object 1 Object N - semantic … -semantic -geometry -geometry Computer vision Image/ video Object 1 Object N - semantic … -semantic -geometry -geometry spatial & temporal relations Computer vision Image/ video Object 1 Object N - semantic … -semantic -geometry -geometry spatial & temporal relations Scene -Sema ntic - geometry Computer vision • Informatio n extraction Sensing device Computational • Interpretation device 1. Information extraction: features, 3D structure, motion flows, etc… 2. Interpretation: recognize objects, scenes, actions, events Computer vision and Applications EosSystems 1990 2000 2010 21 Fingerprint biometrics Augmentation with 3D computer graphics 23 3D object prototyping EosSystems Photomodeler 24 Computer vision and Applications EosSystems Autostich 1990 2000 2010 25 Face detection Face detection Web applications Photometria 28 Panoramic Photography kolor 3D modeling of landmarks 30 Computer vision and Applications • Efficient SLAM/SFM • Large scale image repositories • Deep learning (e.g.
    [Show full text]
  • Bayesian Reconstruction of 3D Human Motion from Single-Camera Video
    Bayesian Reconstruction of 3D Human Motion from Single-Camera Video Nicholas R. Howe Michael E. Leventon Department of Computer Science Artificial Intelligence Lab Cornell University Massachusetts Institute of Technology Ithaca, NY 14850 Cambridge, MA 02139 [email protected] [email protected] William T. Freeman MERL - a Mitsubishi Electric Research Lab 201 Broadway Cambridge, MA 02139 [email protected] Abstract The three-dimensional motion of humans is underdetermined when the observation is limited to a single camera, due to the inherent 3D ambi­ guity of 2D video. We present a system that reconstructs the 3D motion of human subjects from single-camera video, relying on prior knowledge about human motion, learned from training data, to resolve those am­ biguities. After initialization in 2D, the tracking and 3D reconstruction is automatic; we show results for several video sequences. The results show the power of treating 3D body tracking as an inference problem. 1 Introduction We seek to capture the 3D motions of humans from video sequences. The potential appli­ cations are broad, including industrial computer graphics, virtual reality, and improved human-computer interaction. Recent research attention has focused on unencumbered tracking techniques that don't require attaching markers to the subject's body [4, 5], see [12] for a survey. Typically, these methods require simultaneous views from multiple cam­ eras. Motion capture from a single camera is important for several reasons. First, though under­ determined, it is a problem people can solve easily, as anyone viewing a dancer in a movie can confirm. Single camera shots are the most convenient to obtain, and, of course, apply to the world's film and video archives.
    [Show full text]
  • Arxiv:2001.05613V2 [Cs.CV] 14 Oct 2020 Mental Results Demonstrate That the Mean Per Joint Position I.E., Parts Or All of the Body Must Not Be Lost at Any Time
    Synergetic Reconstruction from 2D Pose and 3D Motion for Wide-Space Multi-Person Video Motion Capture in the Wild Takuya Ohashi1,2 Yosuke Ikegami2 Yoshihiko Nakamura2 1NTT DOCOMO 2The University of Tokyo [email protected] [email protected] [email protected] Figure 1: All futsal players’ motions were captured using 12 video cameras surrounding the court. (left) Input images and reprojected joint position. (right) Bone CG drawing based on the calculated joint angles. Abstract diagnosis, behavioral understanding, and even humanoid robot operation [43, 32, 37]. Various motion capture meth- Although many studies have investigated markerless mo- ods have been developed to obtain such data, e.g., opti- tion capture, the technology has not been applied to real cal motion capture, where reflective markers are attached sports or concerts. In this paper, we propose a marker- to characteristic parts of the body, and these 3D positions less motion capture method with spatiotemporal accuracy are then measured [1,5]. Inertial motion capture uses IMU and smoothness from multiple cameras in wide-space and sensors attached to body parts, and then, the positions are multi-person environments. The proposed method predicts calculated using sensor speed [6,2]. Markerless motion each person’s 3D pose and determines the bounding box capture uses a depth camera or single/multiple RGB video of multi-camera images small enough. This prediction and cameras [34, 38,3,4]. However, although various methods spatiotemporal filtering based on human skeletal model en- for using motion data exist, this technology is only used in ables 3D reconstruction of the person and demonstrates limited locations.
    [Show full text]