Contextual information based network with high-frequency feature fusion for high frame rate and ultra-low delay small-scale object detection

Dongmei Huang*, Jihan Zhang, Tingting Hu, Ryuji Fuchikami, Takashi Ikenaga

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

High frame rate and ultra-low delay small-scale object detection plays an important role in factory automation for its timely and accurate reaction. Although many CNN based detection methods have been proposed to improve the accuracy of small object detection for the low resolution and large gap between the object and the background, it is difficult to achieve a trade-off between accuracy and speed. For the pursuit of ultra-low delay processing by utilizing FPGA, this paper proposes: (A) IoU and distance based loss function, (B) Contextual information with high temporal correlation based parallel detection, (C) High frequency feature fusion for enhancing low-bit networks. The proposed methods achieve 45.3 % mAP for test sequences, which is only 0.7 % mAP lower compared with the general method. Meanwhile, the size of the model has been compressed to 1.94 % of the original size and reaches a speed of 278 fPs on FPGA and 15 fPs on GPU.

Original languageEnglish
Title of host publicationProceedings of MVA 2021 - 17th International Conference on Machine Vision Applications
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9784901122207
DOIs
Publication statusPublished - 2021 Jul 25
Event17th International Conference on Machine Vision Applications, MVA 2021 - Aichi, Japan
Duration: 2021 Jul 252021 Jul 27

Publication series

NameProceedings of MVA 2021 - 17th International Conference on Machine Vision Applications

Conference

Conference17th International Conference on Machine Vision Applications, MVA 2021
Country/TerritoryJapan
CityAichi
Period21/7/2521/7/27

ASJC Scopus subject areas

  • Computer Science Applications
  • Computer Vision and Pattern Recognition
  • Signal Processing

Fingerprint

Dive into the research topics of 'Contextual information based network with high-frequency feature fusion for high frame rate and ultra-low delay small-scale object detection'. Together they form a unique fingerprint.

Cite this