Compact Descriptors for Video Analysis: The Emerging MPEG Standard

Duan, Ling-Yu; Lou, Yihang; Bai, Yan; Huang, Tiejun; Gao, Wen; Chandrasekhar, Vijay; Lin, Jie; Wang, Shiqi; Kot, Alex C.

doi:10.1109/mmul.2018.2873844

Cited by 72 publications

(42 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Considering a parking space sharing system (we have its development in §III), surveillance cameras are the sensing devices. To detect whether the sparking lots are occupied and analyze the occupancy history, certain meaningful descriptors [5] will be utilized, in stead of raw video frames.…”

Section: Kcn Frameworkmentioning

confidence: 99%

See 1 more Smart Citation

Vision and Challenges for Knowledge Centric Networking

Wang

et al. 2019

IEEE Wireless Commun.

View full text Add to dashboard Cite

In the creation of a smart future information society, Internet of Things (IoT) and Content Centric Networking (CCN) break two key barriers for both the front-end sensing and back-end networking. However, we still observe the missing piece of the research that dominates the current networking traffic control and system management, e.g., lacking of the knowledge penetrated into both sensing and networking to glue them holistically. In this paper, we envision to leverage emerging machine learning or deep learning techniques to create aspects of knowledge for facilitating the designs. In particular, we can extract knowledge from collected data to facilitate reduced data volume, enhanced system intelligence and interactivity, improved service quality, communication with better controllability and lower cost. We name such a knowledge-oriented traffic control and networking management paradigm as the Knowledge Centric Networking (KCN). This paper presents KCN rationale, KCN benefits, related works and research opportunities.

show abstract

Section: Kcn Frameworkmentioning

confidence: 99%

“…Another active area to leverage deep learning for the knowledge creation is the video processing for Compact Descriptors for Video Analysis (CDVA) [5]. As consecutive frames share many redundancies, in CDVA, certain descriptor, like scalable compressed Fisher Vector (SCFV) or color histogram, is first derived from each frame.…”

Section: Modelsmentioning

confidence: 99%

Vision and Challenges for Knowledge Centric Networking

Wang

et al. 2019

IEEE Wireless Commun.

View full text Add to dashboard Cite

show abstract

“…where N qi and N ri (i from 1 to l) denote the counts of query and reference inlier points, contained in Eq. (13), which fall into the i th bins of H(Q) and H(R), respectively. We know that histogram intersection effectively counts the number of points in two sets that fall into the same bin, and we use this criteria to evaluate the matching degree of H(Q) and H(R).…”

Section: S Lds S Computing Based On Tf-idf Weighted Histogram Matchingmentioning

confidence: 99%

“…To partly address these challenges, the ISO/IEC moving pictures experts group (MPEG) drafts the compact descriptors for visual Manuscript search (CDVS) [10]. To future support the large-scale video retrieval applications, the MPEG is carrying out the draft of Compact Descriptors for Video Analysis (CDVA) [13]. The recent work [14] shows that the combination of hand-crafted feature or descriptor, such as CDVS, and deep features can improve image retrieval performance dramatically.…”

Section: Introductionmentioning

confidence: 99%

Highly Efficient Mobile Visual Search Algorithm

Zhu

Huang

Guo

et al. 2018

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

In this paper, we propose a highly efficient mobile visual search algorithm. For descriptor extraction process, we propose a low complexity feature detection which utilizes the detected local key points of the coarse octaves to guide the scale space construction and feature detection in the fine octave. The Gaussian and Laplacian operations are skipped for the unimportant area, and thus the computing time is saved. Besides, feature selection is placed before orientation computing to further reduce the complexity of feature detection by pre-discarding some unimportant local points. For the image retrieval process, we design a highperformance reranking method, which merges both the global descriptor matching score and the local descriptor similarity score (LDSS). In the calculating of LDSS, the tf-idf weighted histogram matching is performed to integrate the statistical information of the database. The results show that the proposed highly efficient approach achieves comparable performance with the state-of-the-art for mobile visual search, while the descriptor extraction complexity is largely reduced. key words: mobile visual search, descriptor extraction, feature selection, reranking

show abstract

“…Therefore, some feature coding methods that aim to compress and transmit different kinds of extracted features of videos are proposed recently. Duan et al [5] describe the compact descriptors for video analysis, where handcrafted and deep features are compressed and transmitted in a standardized bitstream. Chen et al [6] introduce their proposed Region-of-Interest (ROI) location coding * Recently, reliable human skeletons can be obtained from the depth sensor using real-time skeleton estimation algorithms.…”

Section: Introduction and Related Workmentioning

confidence: 99%

A Multimodal Lossless Coding Method for Skeletons in Videos

Liu

Lin

et al. 2019

2019 IEEE International Conference on Multimedia &Amp; Expo Workshops (ICMEW)

View full text Add to dashboard Cite

Nowadays, skeleton information in videos plays an important role in human-centric video analysis but effective coding such massive skeleton information has never been addressed in previous work. In this paper, we make the first attempt to solve this problem by proposing a multimodal skeleton coding tool containing three different coding schemes, namely, spatial differential-coding scheme, motionvector-based differential-coding scheme and inter prediction scheme, thus utilizing both spatial and temporal redundancy to losslessly compress skeleton data. More importantly, these schemes are switched properly for different types of skeletons in video frames, hence achieving further improvement of compression rate. Experimental results show that our approach leads to 74.4% and 54.7% size reduction on our surveillance sequences and overall test sequences respectively, which demonstrates the effectiveness of our skeleton coding tool.

show abstract

Compact Descriptors for Video Analysis: The Emerging MPEG Standard

Cited by 72 publications

References 7 publications

Vision and Challenges for Knowledge Centric Networking

Vision and Challenges for Knowledge Centric Networking

Highly Efficient Mobile Visual Search Algorithm

A Multimodal Lossless Coding Method for Skeletons in Videos

Contact Info

Product

Resources

About