DOCSLIB.ORG

Unifying Global-Local Representations in Salient Object Detection with Transformer

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Unifying Global-Local Representations in Salient Object Detection with Transformer Unifying Global-Local Representations in Salient Object Detection with Transformer Sucheng Ren1, Qiang Wen1, Nanxuan Zhao2, Guoqiang Han1, Shengfeng He1* 1 School of Computer Science and Engineering, South China University of Technology 2 The Chinese University of Hong Kong Abstract The fully convolutional network (FCN) has dominated salient object detection for a long period. However, the lo- cality of CNN requires the model deep enough to have a global receptive field and such a deep model always leads to the loss of local details. In this paper, we introduce a new attention-based encoder, vision transformer, into salient ob- ject detection to ensure the globalization of the represen- tations from shallow to deep layers. With the global view in very shallow layers, the transformer encoder preserves more local representations to recover the spatial details in final saliency maps. Besides, as each layer can cap- ture a global view of its previous layer, adjacent layers can implicitly maximize the representation differences and Input GT Ours GateNet minimize the redundant features, making that every output feature of transformer layers contributes uniquely for fi- Figure 1: Examples of our method. We propose the Global- nal prediction. To decode features from the transformer, Local Saliency TRansformer (GLSTR) to unify global and we propose a simple yet effective deeply-transformed de- local features in each layer. We compare our method with coder. The decoder densely decodes and upsamples the a SOTA method, GateNet [59], based on FCN architecture. transformer features, generating the final saliency map with Our method can localize salient region precisely with accu- less noise injection. Experimental results demonstrate that rate boundary. our method significantly outperforms other FCN-based and transformer-based methods in five benchmarks by a large and hand-crafted features like color contrast [6], bright- margin, with an average of 12.17% improvement in terms ness [11], foreground and background distinctness [15]. of Mean Absolute Error (MAE). Code will be available at However, these methods limit the representation power, https://github.com/OliverRensu/GLSTR. arXiv:2108.02759v1 [cs.CV] 5 Aug 2021 without considering semantic information. Thus, they of- ten fail to deal with complex scenes. With the development of deep convolutional neural network (CNN), fully convo- 1. Introduction lutional network (FCN) [24] becomes an essential building block for SOD [20, 39, 40, 51]. These methods can encode Salient object detection (SOD) aims at segmenting the high-level semantic features and better localize the salient most visually attracting objects of the images, aligning regions. with the human perception. It gained broad attention in However, due to the nature of locality, methods relying recent years due to its fundamental role in many vision on FCN usually face a trade-off between capturing global tasks, such as image manipulation [27], semantic segmen- and local features. To encode high-level global represen- tation [25], autonomous navigation [7], and person re- tations, the model needs to take a stack of convolutional identification [49, 58]. layers to obtain larger receptive fields, but it erases local de- Traditional methods mainly rely on different priors tails. To retain the local understanding of saliency, features *Corresponding author ([email protected]). from shallow layers fail to incorporate high-level semantics. still preserve rich local and global features. To the best of our knowledge, we are the first work to apply the pure transformer-based encoder on SOD tasks. To sum up, our contributions are in three folds: • We unify global-local representations with a novel transformer-based architecture, which models long- range dependency within each layer. • To take full advantage of the global information of Input GT Layer 1 Layer 12 previous layers, we propose a new decoder, deeply- Figure 2: The visual attention map of the red block in the in- transformed decoder, to densely decode features of put image on the first and twelfth layers of the transformer. each layer. The attention maps show that features from shallow layer can also have global information and features from deep • We conduct extensive evaluations on five widely-used layer can also have local information. benchmark datasets, showing that our method outper- forms state-of-the-art methods by a large margin, with an average of 12:17% improvement in terms of MAE. This discrepancy may make the fusion of global and local features less efficient. Is it possible to unify global and local 2. Related Work features in each layer to reduce the ambiguity and enhance the accuracy for SOD? Salient Object Detection. FCN has became the main- In this paper, we aim to jointly learn global and local stream architecture for the saliency detection methods, to features in a layer-wise manner for solving the salient ob- predict pixel-wise saliency maps directly [14, 20, 23, 31, 36, ject detection task. Rather than using a pure CNN archi- 39, 40, 51, 55]. Especially, the skip-connection architecture tecture, we turn to transformer [35] for help. We get in- has been demonstrated its performance in combining global spired from the recent success of transformer on various and local context information [16,22,26,34,41,52,59]. Hou vision tasks [9] for its superiority in exploring long-range et al. [16] proposed that the high-level features are capable dependency. Transformer applies self-attention mechanism of localizing the salient regions while the shallower ones are for each layer to learn a global representation. Therefore, active at preserving low-level details. Therefore, they intro- it is possible to inject globality in the shallow layers, while duced short connections between the deeper and shallower maintaining local features. As illustrated in Fig.2, the atten- feature pairs. Based on the U-Shape architecture, Zhao et tion map of the red block has the global view of the chicken al. [59] introduced a transition layer and a gate into every and the egg even in the first layer. The network does at- skip connection, to suppress the misleading convolutional tend to small scale details even in the final layer (the twelfth information in the original features from the encoder. In layer). More importantly, with the self-attention mecha- order to localize the salient objects better, they also built a nism, the transformer is capable to model the “contrast”, Fold-ASPP module on the top of the encoder. Similarly, Liu which has demonstrated to be crucial for saliency percep- et al. [22] inserted a pyramid pooling module on the top of tion [8, 28, 32]. the encoder to capture the high-level semantic information. With the merits mentioned above, we propose a method In order to recover the diluted information in the convolu- called Global-Local Saliency TRansformer (GLSTR). The tional encoder and alleviate the coarse-level feature aggre- core of our method is a pure transformer-based encoder and gation problem while decoding, they introduced a feature a mixed decoder to aggregate features generated by trans- aggregation module after every skip connection. In [57], formers. To encode features through transformers, we first Zhao et al. adopted the VGG network without fully con- split the input image into a grid of fixed-size patches. We nected layers as their backbone. They fused the high-level use a linear projection layer to map each image patch to feature into the shallower one at each scale to recover the a feature vector for representing local details. By passing diluted location information. After three convolution lay- the feature through the multi-head self-attention module in ers, each fused feature was enhanced by the saliency map. each transformer layer, the model further encodes global Specifically, they introduced the edge map as a guidance features without diluting the local ones. To decode the fea- for the shallowest feature and fused this feature into deeper tures with global-local information over the inputs and the ones to produce the side outputs. previous layer from the transformer encoder, we propose to Aforementioned methods usually suffer from the lim- densely decode feature from each transformer layer. With ited size of convolutional operation. Although there will this dense connection, the representations during decoding be large receptive fields in the deep convolution layers, the Patch Position Reshape, Pixel Shuffle (x4) Reshape, Pixel Shuffle (x2) Reshape, Conv Layer Embedding Embedding Bilinear Interpolation (x2) Bilinear Interpolation (x2) Bilinear Interpolation (x2) Pixel Shuffle (x2) Transformer Encoder N=16x16 Linear N=16x16 Transformer Transformer Transformer Transformer Projection Layer Layer Layer Layer MLP N=12 Layer Norm Multi-Head Attention Supervision N=4 N=4 H/8 H/2 H/4 Layer Norm H W/8 W/4 Stage 3 W/2 Stage 2 Stage 1 Saliency GT Network output W Deeply-transformed Decoder Figure 3: The pipeline of our proposed method. We first divide the image into non-overlap patches and map each patch into the token before feeding to transformer layers. After encoding features through 12 transformer layers, we decode each output feature in three successive stages with 8×, 4×, and 2× upsampling respectively. Each decoding stage contains four layers and the input of each layer comes from the features of its previous layer together with the corresponding transformer layer. low-level information is diluted increasingly with the in- representations jointly in each encoder layer and decodes creasing number of layers. all features from transformers with gradual spatial resolu- Transformer. Since the local information covered by tion recovery. In the following subsections, we first review the convolutional operation is deficient, numerous methods the local and global feature extraction by the fully convo- adopt attention mechanisms to capture the long-range cor- lutional network-based salient object detection methods in relation [1, 17, 37, 43, 56].
Load more

Recommended publications
  • The Heart of Scotland
    TH E H EART OF S COTLA N D PAINT ED BY SUTTO N PALM ER DESCRIBED BY O PE M CRIEF F A . R H N . O PUBLI SH ED BY 4. SO H O SQUARE ° W A A 59 CH ARLES O O . D M L ND N , BLAC K MCMI " Prefa c e “ BO NNI E SCOTLAND pleased so many readers that it came to be supplemented by another volume dwelling “ ” a the a an d an w m inly on western Highl nds Isl ds, hich was illustrated in a different style to match their Wilder ’ t n the a t a nd mistier features . Such an addi io gave u hor s likeness of Scotland a somewhat lop- sided effect and to balance this list he has prepared a third volume dealing w t a nd e et no t u i h the trimmer rich r, y less pict resque te t t —t at t region of nes visi ed by strangers h is, Per hshire and its e to the Hea rt o bord rs . This is shown be f S cotla nd n as a n t a o , not o ly cont i ing its mos f m us scenery, t e n H d a but as bes bl ndi g ighlan and Lowl nd charms, a nd as having made a focus of the national life and t t t his ory . Pic and Scot, Cel and Sassenach, king and vassal, mailed baron and plaided chief, cateran and farmer, t and n and Jacobi e Hanoveria , gauger smuggler, Kirk a nd e n n o n S cessio , here in tur carried a series of struggles whose incidents should be well known through the ‘ Wa verley Novels .
    [Show full text]
  • Health Beat Issue No. 63
    HEALTH exam Make the Healthier Choice _____ 1. The rubella virus is the virus that causes... a) Chickenpox b) German Measles b) Measles _____ 2. Exclusive breastfeeding means giving only breast milk for babies from the first hour of life up to... a) 4 months old b) 6 months old c) 2 years old _____ 3. Which of the following is considered a dispensable organ or can be safely removed without compromising one’s life... a) Brain c) Heart c) Kidney _____ 4. The most common form of diabetes is called... a) Type 1 Diabetes b) Type 2 Diabetes c) Gestational Diabetes _____ 5. The most common type of childhood cancer in the Philippines is... a) Brain Cancer b) Leukemia c) Lung Cancer _____ 6. The most common man-made source of ionizing radiation that people can be exposed to today is from... a) Cellular Sites b) Nuclear Power Plants c) X-ray Machines _____ 7. The electronic cigarette emits... a) Air b) Smoke c) Vapor _____ 8. To prescribe regulated drugs like morphine, Filipino doctors need... a) Business Permit b) PRC License c) S2 License _____ 9. ISO is not an abbreviation of International Organization for Standardization but derived from the Greek word “isos” meaning... a) Equal b) Partner c) Standard _____ 10. The suffix “cidal” in ovicidal and larvicidal (OL) mosquito traps, a device designed to reduce the population of the dengue-carrying mosquitoes, connotes... a) Catch b) Death c) Hatch Answers on Page 49 March - April 2011 I HEALTHbeat 3 DEPARTMENT OF HEALTH - National Center for Health Promotion 2F Bldg.
    [Show full text]
  • LIHI at Home
    LIHIat home 2013 Ernestine Anderson Place Grand Opening! On February 8th, LIHI held the Grand and a classroom. Opening of Ernestine Anderson Place in Ernestine Anderson Place is named Seattle’s Central District. Ernestine Anderson in honor of legendary jazz singer Ernestine Place features Anderson, an international star from permanent Seattle’s Central Area and graduate of housing linked Garfield High. In a career spanning more with supportive than five decades, she has recorded over services for 45 30 albums. She has sung at Carnegie Hall, homeless seniors, the Kennedy Center, the Monterey Jazz including 17 Festival, as well as at jazz festivals all over veterans, as well the world. as housing for 15 Ernestine Anderson Place is built green low-income seniors. according to Washington State Evergreen The building Standards and designed with long term features generous durability as a priority. The building community space features energy efficient insulation and fan for residents, systems and uses low VOC materials, including a library Energy-Star appliances, and dual-flush with computers, toilets throughout. Ernestine Anderson an exercise room, Place is a non-smoking facility, helping continued on page 3 Vote Each Day in May and Help LIHI Win $250,000 to House Homeless Veterans! Vote daily in May to help us house experiencing homelessness in King County veterans. You can help LIHI win $250,000 alone. With your votes, we’ll win $250,000 to House Homeless Veterans by voting to help build and rehab 50 apartments to assist local Veterans transitioning out of Veteran Thomas English and son live at on Facebook.
    [Show full text]
  • 1 Illl'll Illilllili Lilli 1 1 Illi AZ15022-14 AMC434518-AMC435476
    NOTICE!! These documents have been scanned! Do not place un-scanned documents beneath this notice! Do not remove this notice from this file! GPO Jacket No. 560-102 Print Order 61540 Rise Business Services, LLC job=AZ15 5/6/2019 11'll'll IIIIIM ililli Il ilill ll'lll'llill Box Number= AZ15022 1 lillill 11111 1111111111111111 lil Il l IN Ill'I ll ~ 1111111111 lil ill 11111 lili 11111111 Ill 11111 Ill lil Claim Begin-End: AMC434570-AMC434572 1 Initial Receipt 1 Illl'll Illilllili lilli 111111 11 Illi AZ15022-14 AMC434518-AMC435476 AMC# DATE CLOSED REMARKS 43V Slb 434 51 I 434 <73 1-1 1 United States Department of the Interior Bureau of Land Management Receipt LANDS/RECREATION & PLANNING ONE N CENTRAL AVE PHOENIX, AZ 85004 -2203 No: 3428166 Phone: 602-417-9200 Transaction #: 3527202 Date of Transaction: 11/09/2015 CUSTOMER: MICHEAL LEU 7705 SAINT JAMES DR MENTOR,OH 44060-3988 US g A LINE UNIT # QTY DESCRIPTION REMARKS TOTAL PRICE LOCATABLE MINERALS / MINING CLAIMS- NEW,UNADJUD, ONE OR MORE AUTH NOS / 1 3.00 NEW MINING CLAIMS LOCATION FEE -n/a - 111.00 CASES: AMC434570/$37.00, AMC434571/ $37.00, AMC434572/$37.00 LOCATABLE MINERALS / MINING CLAIMS- NEW,UNADJUD, ONE OR MORE AUTH NOS / 2 3.00 NEW MINING CLAIM PROCESSING FEE -n/a - 60.00 CASES: AMC434570/$20.00, AMC434571/ 520_00, AMC434572/$20.00 LOCATABLE MINERALS / MINING CLAIMS- NEW,UNADJUD, ONE OR MORE AUTH NOS / 3 3.00 NEW MINING CLAIMS MAINTENANCE FEE -n/a - 465.00 CASES: AMC434570/$155.00, AMC434571/ $155.00, AMC434572/$155.00 LOCATABLE MINERALS / MINING CLAIMS- NOT NEW-UNADJUD,ONE AUTH NO.
    [Show full text]
  • Arxiv:2007.09062V1 [Cs.CV] 17 Jul 2020 Most Visually Obvious Regions
    Multi-scale Interactive Network for Salient Object Detection Youwei Pang1y, Xiaoqi Zhao1y, Lihe Zhang1∗and Huchuan Lu1;2 1Dalian University of Technology, China 2Peng Cheng Laboratory flartpang, [email protected], fzhanglihe, [email protected] L Abstract a r g e Deep-learning based salient object detection methods achieve great progress. However, the variable scale and M i d unknown category of salient objects are great challenges all d l the time. These are closely related to the utilization of multi- e level and multi-scale features. In this paper, we propose the aggregate interaction modules to integrate the features from S m a l adjacent levels, in which less noise is introduced because of l only using small up-/down-sampling rates. To obtain more Image GT MINet AFNet CPD PAGR efficient multi-scale features from the integrated features, Figure 1. Several visual examples with size-varying objects and the self-interaction modules are embedded in each decoder their predictions generated by the proposed MINet, AFNet [9], unit. Besides, the class imbalance issue caused by the scale CPD [41] and PAGR [53] methods. variation weakens the effect of the binary cross entropy loss and results in the spatial inconsistency of the predic- tions. Therefore, we exploit the consistency-enhanced loss ods, which are limited by the absence of necessary detailed to highlight the fore-/back-ground difference and preserve information owing to the repeated sub-sampling, have diffi- the intra-class consistency. Experimental results on five culty in consistently and accurately segmenting salient ob- benchmark datasets demonstrate that the proposed method jects of different scales (Fig.1).
    [Show full text]
  • Page 1 DOCUMENT RESUME ED 335 965 FL 019 564 AUTHOR
    DOCUMENT RESUME ED 335 965 FL 019 564 AUTHOR Riego de Rios, Maria Isabelita TITLE A Composite Dictionary of Philippine Creole Spanish (PCS). INSTITUTION Linguistic Society of the Philippines, Manila.; Summer Inst. of Linguistics, Manila (Philippines). REPORT NO ISBN-971-1059-09-6; ISSN-0116-0516 PUB DATE 89 NOTE 218p.; Dissertation, Ateneo de Manila University. The editor of "Studies in Philippine Linguistics" is Fe T. Otanes. The author is a Sister in the R.V.M. order. PUB TYPE Reference Materials - Vocabularies/Classifications/Dictionaries (134)-- Dissertations/Theses - Doctoral Dissertations (041) JOURNAL CIT Studies in Philippine Linguistics; v7 n2 1989 EDRS PRICE MF01/PC09 Plus Postage. DESCRIPTORS *Creoles; Dialect Studies; Dictionaries; English; Foreign Countries; *Language Classification; Language Research; *Language Variation; Linguistic Theory; *Spanish IDENTIFIERS *Cotabato Chabacano; *Philippines ABSTRACT This dictionary is a composite of four Philippine Creole Spanish dialects: Cotabato Chabacano and variants spoken in Ternate, Cavite City, and Zamboanga City. The volume contains 6,542 main lexical entries with corresponding entries with contrasting data from the three other variants. A concludins section summarizes findings of the dialect study that led to the dictionary's writing. Appended materials include a 99-item bibliography and materials related to the structural analysis of the dialects. An index also contains three alphabetical word lists of the variants. The research underlying the dictionary's construction is
    [Show full text]
  • Ocular Side Effects of Novel Anti-Cancer Biological Therapies
    www.nature.com/scientificreports OPEN Ocular side efects of novel anti‑cancer biological therapies Vicktoria Vishnevskia‑Dai1*, Lihi Rozner1, Raanan Berger2,3, Ziv Jaron1, Sivan Elyashiv1, Gal Markel2,3 & Ofra Zloto1 To examine the ocular side efects of selected biological anti‑cancer therapies and the ocular and systemic prognosis of patients receiving them. We retrospectively reviewed all medical records of patients who received biological anti‑cancer treatment from 1/2012 to 12/2017 and who were treated at our ocular oncology service. The following data was retrieved: primary malignancy, metastasis, type of biological therapy, ocular side efects, ophthalmic treatment, non‑ocular side efects, and ocular and systemic disease prognoses. Twenty‑two patients received biological therapies and reported ocular side efects. Eighteen patients (81.8%) had bilateral ocular side efects, including uveitis (40.9%), dry eye (22.7%), and central serous retinopathy (22.7%). One patient (4.5%) had central retinal artery occlusion (CRAO), and one patient (4.5%) had branch retinal vein occlusion (BRVO). At the end of follow‑up, 6 patients (27.27%) had resolution of the ocular disease, 13 patients (59.09%) had stable ocular disease, and 3 patients (13.64%) had progression of the ocular disease. Visual acuity improved signifcantly at the end of follow‑up compared to initial values. Eighteen patients (81.8%) were alive at study closure. Biological therapies can cause a wide range of ocular side efects ranging from dry eye symptoms to severe pathologies that may cause ocular morbidity and vision loss, such as uveitis, CRAO and BRVO. All patients receiving biological treatments should be screened by ophthalmologists before treatment, re‑screened every 4–6 months during treatment, and again at the end of treatment.
    [Show full text]
  • Maternal and Child Care Among the Tagalogs in Bay, Laguna, Philippines
    MATERNAL AND CHILD CARE AMONG. THE .TAGALOGS IN BAY, LAGUNA, PHiliPPINES .. > ' • • • F. LANDA INTRODL'CTION The purpose of this paper· is to present an ethnographic picture of certain aspects of maternal· and child care among the Tagalogs in- habiting the municipality of Bay, Laguna, Philippines.1 It is hoped that persons working in programs of family planning,· maternal ·and child health, and ·community mediCine will find the data useful. No· theoretical model is worked into datiL Our aim is to illustrate empirica1ly that traditional practices · associated with maternal and child care are at an· guesswork: as mbst health innovators often think them. to 'Matermtl and child 'care in the area is hand1ed by individuals who adept practition.ers. Their training and skills differ from those of P;l.qdern .physicians and ilurses, if ·only because the medical technology available. in. the community is less developed than that found in urban centers and. universities.. But. this does I not mean' · that folk medical practices are based entirely on unsound medical knowledge. This asser- 'tion becomes clear if one assesses medical practices in Bay in the context of . the local technology known and accessible to the people. The merit of this. assertion lies on the fact that for a peasant group to be able to develop standardized ways of handlip.g medical problems, to· cultivate wild vines and grasses as effective medicinal plants and abortifacients, to recognize disease and prescribe the best .plant to cure 'it, to formulate a body of beliefs that serves as guideline for systematic medical action - this, to my mind, is enough .
    [Show full text]
  • Adaptive and Attentive Depth Distiller for Efficient RGB-D Salient Object
    A2dele: Adaptive and Attentive Depth Distiller for Efficient RGB-D Salient Object Detection Yongri Piao1∗ Zhengkun Rong1∗ Miao Zhang1,2† Weisong Ren1 Huchuan Lu1,3 1Dalian University of Technology, China 2Key Lab for Ubiquitous Network and Service Software of Liaoning Province, Dalian University of Technology, China 3Pengcheng Lab {yrpiao, miaozhang, lhchuan}@dlut.edu.cn, {rzk911113, beatlescoco}@mail.dlut.edu.cn 0.89 Abstract RGB 0.88 RGB+Depth CPFP'19+A2dele Our Existing state-of-the-art RGB-D salient object detection 0.87 methods explore RGB-D data relying on a two-stream ar- 0.86 DMRA'19 chitecture, in which an independent subnetwork is required 0.85 to process depth data. This inevitably incurs extra compu- DMRA'19+A2dele 0.84 tational costs and memory consumption, and using depth CPFP'19 data during testing may hinder the practical applications of F-measure 0.83 DMRA'19 RGB-D saliency detection. To tackle these two dilemmas, 0.82 we propose a depth distiller (A2dele) to explore the way of 0.81 CPFP'19 using network prediction and attention as two bridges to 0.8 transfer the depth knowledge from the depth stream to the 0.79 50 100 150 200 250 300 RGB stream. First, by adaptively minimizing the differences Model Size / MB between predictions generated from the depth stream and Figure 1. F-measure vs. Model Size on NLPR dataset [30]. By RGB stream, we realize the desired control of pixel-wise embedding our A2dele (CPFP’19 [41]+A2dele and DMRA’19 depth knowledge transferred to the RGB stream.
    [Show full text]
  • 10.5.5.2 Final Report of Subanen Zamboanga Del Sur Size
    Phase II Documentation of Philippine Traditional Knowledge and Practices on Health and Development of Traditional Knowledge Digital Library on Health for Selected Ethnolinguistic Groups: The SUBANEN people of Salambuyan, Lapuyan, Zamboanga del Sur REPORT PREPARED BY: Marilou C. Elago, Western Mindanao State University, Zamboanga City Rhea Felise A. Dando, University of the Philippines Manila, Ermita, Manila Jhoan Rhea L. Pizon, Western Mindanao State University, Zamboanga City Rainier M. Galang, University of the Philippines Manila, Ermita, Manila Isidro C. Sia, University of the Philippines Manila, Ermita, Manila 2013 Summary An ethnopharmacological study of the Subanen was conducted from May 2012 to May 2013. The one-year study included documentation primarily of the indigenous healing practices and ethnopharmacological knowledge of the Subanen. The ethnohistorical background of the tribe was also included in the study. The study covered the Barangay Salambuyan, Zamboanga del Norte. A total of 132 plants and 3 other natural products, 7 traditional healers and 3 focus group discussions in the community were documented. Documentation employed the use of prepared ethnopharmacological templates which include: medicinal plants and other natural products, herbarial compendium of selected medicinal plants, local terminology of condition and treatments, rituals and practices, and traditional healer’s templates. Immersion in the community was the primary method employed. Interview and participant- observation, and forest visits were utilized to gather data. Focus group discussions were also done as a form of data validation. Formalized informed consent for this study was asked from National Commission of Indigenous People, barangay officials, and from different key individuals prior to the documentation and collection of medicinal plants.
    [Show full text]
  • Newton Falls Recertification Application Appendices
    ATTACHMENT A QUESTION 3: PROJECT MAP AERIAL PHOTOS St. Lawrence County, New York ± Clinton Franklin St. Lawrence Jefferson ^ Essex Lewis Herkimer Hamilton PROJECT LOCATION Browns Falls MAP FERC Project No. 2713 Newton Falls d x !( Hydroelectric Project m . p a !( !( FERC No. 7000 M n o i t a c o L _ I Upper Newton Falls H I L \ FERC Project No. 7000 Legend s c o d !( Newton Falls Project (P-7000) _ Lower Newton Falls p a !( Other Oswegatchie River Dams m FERC Project No. 7000 \ e c n a i l p m o C t s e W 0 2,000 4,000 6,000 Y N Feet _ 1 0 3 3 3 1 \ d l e i f k o o r B \ s t c e j o r P \ : G : h t a P Data Source: USGS Quadragle 1:24,00; Newton Falls and Oswegatchie UPPER NEWTON FALLS DEVELOPMENT LOWER NEWTON FALLS DEVELOPMENT ATTACHMENT B QUESTION 6: JULY 15, 2002 NEWTON FALLS PROJECT SETTLEMENT OFFER AUGUST 13, 2003 ORDER ISSUING NEW LICENSE (P-7000) DECEMBER 20, 2002 WATER QUALITY CERTIFICATION Jnofflclal FERC-Generated PDF of 20030106-0375 Received by FERC OSEC 12/20/2002 in Docket#: P-7000-000 • New York State Department of Environmental Conservation Dlvlslon of Envlronmental Permits, Region 6 Dulles State Office Building, 317 Washington S~'eet, Watertown, New York 13601-3787 Phone: (315) 785-2245 • FAX: (315) 785-2242 Website: F-~ M. Crot~ ORIGINAL Commmio~r December 20, 2002 Samuel S. Hirschey, Manager Hydro Licensing and Regulatory Compliance Orion Power New York GP 11, Inc.
    [Show full text]
  • Executive Director Monthly Update Council Passes South
    Tweet Share this Page: MAY 2011 Executive Director Monthly Executive Director Monthly Update Update Last week we held HDC's Third Annual Luncheon. As always... Last week we held HDC’s Third Annual Read More Luncheon. As always it was an inspiring event from the standpoint of the attendees alone: 400 of the most skilled and experienced housers in Council Passes South Downtown the state and their supporters including three Rezone On April 25, Seattle City Council Leadership Development Program cohorts, a unanimously approved... slew of local elected officials and Read More representatives from HUD, for both of our state’s U.S. Senators and a Congressman with almost 50 repeat or new sponsors. The event was a Recap of King County Mayors success from that perspective alone and also Panel raised about $100,000 for continued HDC operations in advocacy Last month Kent Mayor Suzette Cook, Issaquah Mayor Ava... and member services. My sincere and profuse thanks to all who Read More joined us that afternoon and to all who contributed as individuals and as sponsors! HDC May Member Meeting: Public Our Luncheon keynote speaker Shelley Poticha, the HUD Director Funders Panel for Sustainable Communities initiatives, spoke highly of our region Join us on May 13th for our annual and its work in linking affordable housing, lower-cost transportation Public Funders Panel... Read More options and environmental stewardship into a single development concept that creates healthy, livable places to live, work and play. HDC members were moving in that direction even before this Member Highlight: Low Income initiative was launched and are now embracing these concepts in Housing Institute their newest projects.
    [Show full text]