2018
DOI: 10.1109/tip.2018.2794203
|View full text |Cite
|
Sign up to set email alerts
|

Fast MPEG-CDVS Encoder With GPU-CPU Hybrid Computing

Abstract: The compact descriptors for visual search (CDVS) standard from ISO/IEC moving pictures experts group has succeeded in enabling the interoperability for efficient and effective image retrieval by standardizing the bitstream syntax of compact feature descriptors. However, the intensive computation of a CDVS encoder unfortunately hinders its widely deployment in industry for large-scale visual search. In this paper, we revisit the merits of low complexity design of CDVS core techniques and present a very fast CDV… Show more

Help me understand this report
View preprint versions

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

0
6
0

Year Published

2018
2018
2022
2022

Publication Types

Select...
9

Relationship

0
9

Authors

Journals

citations
Cited by 14 publications
(6 citation statements)
references
References 35 publications
0
6
0
Order By: Relevance
“…CDVA further extends the concept CDVS by including deep features alongside the handcrafted ones and exploiting temporal redundancies between video frames. We found only three real-time implementations of CDVS [43,51,133] reaching 98, 144 and 39 megapixels per second on desktop-grade GPUs, respectively. Compared to other implementations reported later in the survey, the results are relatively slow, and more research would be necessary to decrease the computational complexity of CDVS and CDVA.…”
Section: The Role Of Machine Learningmentioning
confidence: 90%
“…CDVA further extends the concept CDVS by including deep features alongside the handcrafted ones and exploiting temporal redundancies between video frames. We found only three real-time implementations of CDVS [43,51,133] reaching 98, 144 and 39 megapixels per second on desktop-grade GPUs, respectively. Compared to other implementations reported later in the survey, the results are relatively slow, and more research would be necessary to decrease the computational complexity of CDVS and CDVA.…”
Section: The Role Of Machine Learningmentioning
confidence: 90%
“…General purpose graphical processor units (GPGPU) are becoming more and more invasive in every-day life [13][14][15][16][17][18].…”
Section: Gpgpu and Parallel Programming Basic Notionsmentioning
confidence: 99%
“…Even if we improve the communication performance of the network, the lossy compression part will become the bottleneck to transfer the video stream to the abnormal detection part due to the heavy calculation of the compression processes. To overcome the difficulty of improving the lossy compression performance, we can accelerate the compressor and the decompressor by applying GPUs [ 15 ] and dedicated hardware [ 16 ]. However, the lossy compression reduces information of pixels in frames and guarantees the bandwidth of the video stream.…”
Section: Introductionmentioning
confidence: 99%