Internet Archive Scholar
Filters
























19 Hits in 3.2 sec

Fast MPEG-CDVS Encoder With GPU-CPU Hybrid Computing

Ling-Yu Duan, Wei Sun, Xinfeng Zhang, Shiqi Wang, Jie Chen, Jianxiong Yin, Simon See, Tiejun Huang, Alex C. Kot, Wen Gao
2018 IEEE Transactions on Image Processing  

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (5.9 MB)
https://web.archive.org/web/20200928093654/https://arxiv.org/pdf/1705.09776v1.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.

Note that this fulltext copy is not of the "primary" version of this work. The version it corresponds to is:

2017 pre-print arXiv:1705.09776v1 fatcat:klvpzvzdlffrlgekq6wxklh7f4

Other Versions

2017 accepted
arXiv:1705.09776v2 fatcat:bxgh7t5j2jblphthlirfuozsny
In this paper, we revisit the merits of low complexity design of CDVS core techniques and present a very fast CDVS encoder by leveraging the massive parallel execution resources of GPU.  ...  Comprehensive experimental results over benchmarks are evaluated, which has shown that the fast CDVS encoder using GPU-CPU hybrid computing is promising for scalable visual search.  ...  The fast CDVS encoder is implemented based on the latest CDVS reference software TM14.0 using GPU-CPU hybrid computing.  ... 
doi:10.1109/tip.2018.2794203 pmid:29432101 fatcat:2dr3sc66djcwnofwht6be2n7zm
Multiple Versions
Citation

Ling-Yu Duan, Wei Sun, Xinfeng Zhang, Shiqi Wang, Jie Chen, Jianxiong Yin, Simon See, Tiejun Huang, Alex C. Kot, Wen Gao. "Fast MPEG-CDVS Encoder With GPU-CPU Hybrid Computing." IEEE Transactions on Image Processing 27.5 (2018) 2201-2216

Accelerating Local Feature Extraction Using Two Stage Feature Selection and Partial Gradient Computation [chapter]

Keundong Lee, Seungjae Lee, Weon-Geun Oh
2015 Lecture Notes in Computer Science  

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (1.4 MB)
https://web.archive.org/web/20170829034600/http://vigir.missouri.edu/~gdesouza/Research/Conference_CDs/ACCV_2014/pages/workshop14/pdffiles/w14-p10.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL. The file type is application/pdf.

In this paper, we present a fast local feature extraction method, which is our contribution to ongoing MPEG standardization of compact descriptor for visual search (CDVS).  ...  For its efficiency, the proposed method has been integrated into CDVS TM since 107 th MPEG meeting.  ...  Evaluation on MPEG CDVS Framework The proposed method has been integrated into CDVS TM as an fast mode of feature extractor for its efficiency since 107 th meeting.  ... 
doi:10.1007/978-3-319-16634-6_27 fatcat:5xziekzd2jhktlofe5jbhnxsra
Citation

Keundong Lee, Seungjae Lee, Weon-Geun Oh. "Accelerating Local Feature Extraction Using Two Stage Feature Selection and Partial Gradient Computation." Lecture Notes in Computer Science (2015) 366-380

Compact Global Descriptors for Visual Search

Vijay Chandrasekhar, Jie Lin, Olivier Morere, Antoine Veillard, Hanlin Goh
2015 2015 Data Compression Conference  

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (345.6 kB)
https://web.archive.org/web/20170809145449/http://vijaychan.github.io/Publications/2015%20-%20Compact%20Global%20Descriptors%20for%20Visual%20Search.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL. The file type is application/pdf.

State-of-the-art global descriptors based on Fisher Vectors are represented with tens of thousands of floating point numbers.  ...  Motivated by the remarkable success of deep neural networks in recent literature, we propose a compression scheme based on deeply stacked Restriction Boltzmann Machines (SRBM), which learn lower dimensional  ...  The MPEG-CDVS standard adopted the Scalable Fisher Compressed Vector [6] , which was based on binarization of high-dimensional Fisher Vectors.  ... 
doi:10.1109/dcc.2015.54 dblp:conf/dcc/ChandrasekharLM15 fatcat:mdrrky2pvfd33dvdhc7ih7oage
Citation

Vijay Chandrasekhar, Jie Lin, Olivier Morere, Antoine Veillard, Hanlin Goh. "Compact Global Descriptors for Visual Search." 2015 Data Compression Conference (2015) 333-342

Intermediate Deep Feature Compression: the Next Battlefield of Intelligent Sensing [article]

Zhuo Chen, Weisi Lin, Shiqi Wang, Lingyu Duan, Alex C. Kot
2018 arXiv   pre-print

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (464.5 kB)
https://web.archive.org/web/20200915161009/https://arxiv.org/pdf/1809.06196v1.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.

This strategy enables a good balance among the computational load, transmission load and the generalization ability for cloud servers when deploying the deep neural networks for large scale cloud based  ...  Based on CDVS, MPEG has moved forward to the standardization of Compact Descriptors for Video Analysis (CDVA) [21] since Feb. 2015.  ...  For hand-crafted features, the standards from MPEG including MPEG-CDVS [20] and MPEG-CDVA [21] specify the feature extraction and compression processes.  ... 
arXiv:1809.06196v1 fatcat:qzvvqtyjtvf6zjp2h4xa7r5qbm
Citation

Zhuo Chen, Weisi Lin, Shiqi Wang, Lingyu Duan, Alex C. Kot. "Intermediate Deep Feature Compression: the Next Battlefield of Intelligent Sensing." arXiv (2018)

Tiny Descriptors for Image Retrieval with Unsupervised Triplet Hashing [article]

Jie Lin, Olivier Morère, Julie Petta, Vijay Chandrasekhar, Antoine Veillard
2015 arXiv   pre-print

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (423.8 kB)
https://web.archive.org/web/20200905020316/https://arxiv.org/pdf/1511.03055v1.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.

A good image descriptor is key to the retrieval pipeline and should reconcile two contradictory requirements: providing recall rates as high as possible and being as compact as possible for fast matching  ...  Then, triplet networks, a rank learning scheme based on weight sharing nets is used to fine-tune the binary embedding functions to retain as much as possible of the useful metric properties of the original  ...  The size of the compressed descriptor in the MPEG-CDVS standard ranges from 256 bytes to several thousand bytes per image, based on the operating points.  ... 
arXiv:1511.03055v1 fatcat:owk7tvr3ibectc6f5u2knokggi
Citation

Jie Lin, Olivier Morère, Julie Petta, Vijay Chandrasekhar, Antoine Veillard. "Tiny Descriptors for Image Retrieval with Unsupervised Triplet Hashing." arXiv (2015)

Tiny Descriptors for Image Retrieval with Unsupervised Triplet Hashing

Jie Lin, Olivier Morere, Julie Petta, Vijay Chandrasekhar, Antoine Veillard
2016 2016 Data Compression Conference (DCC)  

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (200.0 kB)
https://web.archive.org/web/20200321104150/http://vijaychan.github.io/Publications/2016_DCC.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.

Further, it is highly desirable that the global descriptors be binary to enable fast matching through Hamming distances.  ...  For the first step, state-of-the-art schemes are based on comparing global representations of images.  ...  The size of the compressed descriptor in the MPEG-CDVS standard ranges from 256 bytes to several thousand bytes per image, based on the operating points.  ... 
doi:10.1109/dcc.2016.23 dblp:conf/dcc/LinMPCV16 fatcat:qxdp7hxcpbe6ffk3ck5jta45pu
Citation

Jie Lin, Olivier Morere, Julie Petta, Vijay Chandrasekhar, Antoine Veillard. "Tiny Descriptors for Image Retrieval with Unsupervised Triplet Hashing." 2016 Data Compression Conference (DCC) (2016) 397-406

AccMPEG: Optimizing Video Encoding for Video Analytics [article]

Kuntai Du, Qizheng Zhang, Anton Arapin, Haodong Wang, Zhengxu Xia, Junchen Jiang
2022 arXiv   pre-print

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (8.2 MB)
https://web.archive.org/web/20220617222500/https://arxiv.org/pdf/2204.12534v1.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL. The file type is application/pdf.

This paper presents AccMPEG, a new video encoding and streaming system that meets all the three requirements.  ...  The key is to learn how much the encoding quality at each (16x16) macroblock can influence the server-side DNN accuracy, which we call accuracy gradient.  ...  Fast training of ONLINE ENCODING We now describe AccMPEG's online encoding process, including the architecture of AccM odel and how AccMPEG assigns encoding quality to each macroblock.  ... 
arXiv:2204.12534v1 fatcat:ou73kl7y25agngtrqhjixggrfu
Open Access
Citation

Kuntai Du, Qizheng Zhang, Anton Arapin, Haodong Wang, Zhengxu Xia, Junchen Jiang. "AccMPEG: Optimizing Video Encoding for Video Analytics." arXiv (2022)

Mobile media communication, processing, and analysis: A review of recent advances

Wen Gao, Ling-Yu Duan, Jun Sun, Junsong Yuan, Yonggang Wen, Yap-Peng Tan, Jianfei Cai, Alex C. Kot
2013 2013 IEEE International Symposium on Circuits and Systems (ISCAS2013)  

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (733.4 kB)
https://web.archive.org/web/20170809013009/http://www.ntu.edu.sg/home/ygwen/Paper/GDS-ISCAS-13.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL. The file type is application/pdf.

To identify the opportunities and challenges in fast growing mobile media computing, we discuss several emerging topics including mobile visual search, retargeting, mobile video streaming, and cloud based  ...  In particular, the ongoing MPEG standardization of compact descriptors for visual search (CDVS) have involved big industry and academia efforts from STMicroelectronics, NEC, Nvidia, Samsung, Nokia, Qualcomm  ...  Hardware acceleration like GPU has been tried out [26] [29] . In [30] , a novel FPGA based architecture was proposed to deal with computational complexity.  ... 
doi:10.1109/iscas.2013.6571985 dblp:conf/iscas/GaoD0YWTCK13 fatcat:mhnjsvlrfvdr3nchgkqhk4i4zi
Citation

Wen Gao, Ling-Yu Duan, Jun Sun, Junsong Yuan, Yonggang Wen, Yap-Peng Tan, Jianfei Cai, Alex C. Kot. "Mobile media communication, processing, and analysis: A review of recent advances." 2013 IEEE International Symposium on Circuits and Systems (ISCAS2013) (2013) 869-872

ACTNET: end-to-end learning of feature activations and multi-stream aggregation for effective instance image retrieval [article]

Syed Sameed Husain, Eng-Jon Ong, Miroslaw Bober
2020 arXiv   pre-print

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (3.6 MB)
https://web.archive.org/web/20201029013817/https://arxiv.org/pdf/1907.05794v3.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.

Other Versions

2019 pre-print
arXiv:1907.05794v2 fatcat:socxcfsbifdnvfcqoiagrgo3gi
2019 pre-print
arXiv:1907.05794v1 fatcat:c4cwkjscjfhc7leudqy5nd45we
Recent works on using activation functions to help fast training of CNNs are [30] , [31] , [32] , [33] , [34] .  ...  We overcome the TITAN X GPU memory limitation of 11 GB by processing one triplet at a time and updating the gradients after every 64 triplets.  ...  Miroslaw led the development of ISO MPEG standards for over 20 years, chairing the MPEG-7, CDVS and CVDA groups. He is an inventor of over 80 patents, many deployed in products.  ... 
arXiv:1907.05794v3 fatcat:ukmpjk53wvgq3brok354qqwmim
Multiple Versions
Citation

Syed Sameed Husain, Eng-Jon Ong, Miroslaw Bober. "ACTNET: end-to-end learning of feature activations and multi-stream aggregation for effective instance image retrieval." arXiv (2020)

Feature extraction using MPEG-CDVS and Deep Learning with application to robotic navigation and image classification

Pedro Porto Buarque De Gusmao, Enrico Magli
2017

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (6.5 MB)
https://web.archive.org/web/20200216112328/https://iris.polito.it/retrieve/handle/11583/2665943/144519/PhDthesisPedroPBGusmao%20%283%29.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.

We first describe a probabilistic approach to loop detection based on the standard's suggested similarity metric.  ...  We then evaluate the performance of CDVS compression modes in terms of matching speed, feature extraction, and storage requirements and compare them with the state of the art SIFT descriptor for five different  ...  This node encapsulates the MPEG-CDVS Test Model's [107] implementation of both CDVS Feature Detector and Extractor. • cdvsMatch: Receives a sequence of MPEG-CDVS bitstreams generated by the cdvsExtract  ... 
doi:10.6092/polito/porto/2665943 fatcat:xe6g2rue5fgphehdetemdhviva
Citation

Pedro Porto Buarque De Gusmao, Enrico Magli. "Feature extraction using MPEG-CDVS and Deep Learning with application to robotic navigation and image classification." (2017)

A survey on compact features for visual content analysis

Luca Baroffio, Alessandro E. C. Redondi, Marco Tagliasacchi, Stefano Tubaro
2016 APSIPA Transactions on Signal and Information Processing  

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (1.0 MB)
https://web.archive.org/web/20190322180124/https://core.ac.uk/download/pdf/55259790.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL. The file type is application/pdf.

features constitute compact yet effective representations of visual content, and are being exploited in a large number of heterogeneous applications, including augmented reality, image registration, content-based  ...  Besides extraction, a large body of research addressed the problem of ad-hoc feature encoding methods, and a number of networking and transmission protocols enabling distributed visual content analysis  ...  In particular, SIFT is widely regarded as the gold standard in the context of local feature extraction, and has been partially adopted by the MPEG Compact Descriptors for Visual Search (CDVS) [4, 130]  ... 
doi:10.1017/atsip.2016.13 fatcat:lokgfydqrrd6zcvgngcbloopbu
DOAJ
Citation

Luca Baroffio, Alessandro E. C. Redondi, Marco Tagliasacchi, Stefano Tubaro. "A survey on compact features for visual content analysis." APSIPA Transactions on Signal and Information Processing (2016)

Digital FPGA Circuits Design for Real-Time Video Processing with Reference to Two Application Scenarios

Giorgio Lopez
2015

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (2.4 MB)
https://web.archive.org/web/20200207180312/http://www.fedoa.unina.it/10491/1/Lopez_Giorgio.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.

General purpose CPU or GPU software implementations of these applications are quite simple and widespread, but commonly do not allow high performance because of the high layering that separates high level  ...  The most practised approach nowadays is based on the use of Very-Large-Scale Integrated (VLSI) digital electronic circuits.  ...  MPEG 7: "the bits about the bits" MPEG 7 is a standard for content description of multimedia formats, so it does not deal with the actual encoding of data streams like other MPEG standards (MPEG-2, MPEG  ... 
doi:10.6092/unina/fedoa/10491 fatcat:4bl5re7ci5dyjhbap3ih2bgqzq
Citation

Giorgio Lopez. "Digital FPGA Circuits Design for Real-Time Video Processing with Reference to Two Application Scenarios." (2015)

Report on Dissemination and Standardisation Activities Y2 [article]

Uwe Riemann
2015

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (661.7 kB)
https://web.archive.org/web/20200211143151/http://digital.joanneum.at/DOI/ICoSOLE-D7.2.2-DTO-DisseminationStandardisationY2-v05.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.

descriptors for visual search (CDVS), which focuses on similarity matching of still images.  ...  on Augmented 360 Degree Video via Gesture-Based Oscillation Compensating Dynamic Adaptive Streaming over HTTP Selecting User Generated Content for Use in Media Productions Interaction A GPU-Accelerated  ... 
doi:10.7800/304icosoled722 fatcat:33zrolhx5rbszp3dpfm5jfmwfe
Citation

Uwe Riemann. "Report on Dissemination and Standardisation Activities Y2." (2015)

HPatches: A benchmark and evaluation of handcrafted and learned local descriptors [article]

Vassileios Balntas and Karel Lenc and Andrea Vedaldi and Krystian Mikolajczyk
2017 arXiv   pre-print

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (1.2 MB)
https://web.archive.org/web/20191017121220/https://arxiv.org/pdf/1704.05939v1.pdf

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL. The file type is application/pdf.

We show that a simple normalisation of traditional hand-crafted descriptors can boost their performance to the level of deep learning based descriptors within a realistic benchmarks evaluation.  ...  The CVDS dataset [9] addresses the data diversity issue by extracting patches from five MPEG-CDVS: Graphics, Paintings, Video, Buildings and Common Objects.  ...  on shallow convolutional networks, triplet learning constraints and fast hard negative mining.  ... 
arXiv:1704.05939v1 fatcat:v2hvvhwomvgzzmmj66n3az46uy
Citation

Vassileios Balntas and Karel Lenc and Andrea Vedaldi and Krystian Mikolajczyk. "HPatches: A benchmark and evaluation of handcrafted and learned local descriptors." arXiv (2017)

HPatches: A Benchmark and Evaluation of Handcrafted and Learned Local Descriptors

Vassileios Balntas, Karel Lenc, Andrea Vedaldi, Krystian Mikolajczyk
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  

Preserved Fulltext

fulltext thumbnail
Web Archive Capture PDF (453.1 kB)
https://web.archive.org/web/20201106164617/https://ora.ox.ac.uk/objects/uuid:76702277-37ab-445c-90e1-9960dad369f8/download_file?safe_filename=Vedaldi%2Bet%2Bal%252C%2BHPatches%2B-%2BA%2Bbenchmark%2Band%2Bevaluation%2Bof%2Bhandcrafted%2Band%2Blearned%2Blocal%2Bdescriptors.pdf&file_format=application%2Fpdf&type_of_work=Conference+item

A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL. The file type is application/pdf.

We show that a simple normalisation of traditional hand-crafted descriptors can boost their performance to the level of deep learning based descriptors within a realistic benchmarks evaluation.  ...  The CVDS dataset [9] addresses the data diversity issue by extracting patches from five MPEG-CDVS: Graphics, Paintings, Video, Buildings and Common Objects.  ...  Line color encodes dataset and line style a detector.  ... 
doi:10.1109/cvpr.2017.410 dblp:conf/cvpr/BalntasLVM17 fatcat:xj2wict7fjhwjkxubpuoeeplem
Citation

Vassileios Balntas, Karel Lenc, Andrea Vedaldi, Krystian Mikolajczyk. "HPatches: A Benchmark and Evaluation of Handcrafted and Learned Local Descriptors." 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2017) 3852-3861

« Previous Showing results 1 — 15 out of 19 results