TY - GEN
T1 - Analysis and a solution of momentarily missed detection for anchor-based object detectors
AU - Hosoya, Yusuke
AU - Suganuma, Masanori
AU - Okatani, Takayuki
N1 - Funding Information:
This work was partly supported by JSPS KAKENHI Grant Number JP15H05919 and JP19H01110 and JST CREST Grant Number JPMJCR14D1.
Publisher Copyright:
© 2020 IEEE.
PY - 2020/3
Y1 - 2020/3
N2 - The employment of convolutional neural networks has led to significant performance improvement on the task of object detection. However, when applying existing detectors to continuous frames in a video, we often encounter momentary miss-detection of objects, that is, objects are undetected exceptionally at a few frames, although they are correctly detected at all other frames. In this paper, we analyze the mechanism of how such miss-detection occurs. For the most popular class of detectors that are based on anchor boxes, we show the followings: i) besides apparent causes such as motion blur, occlusions, background clutters, etc., the majority of remaining miss-detection can be explained by an improper behavior of the detectors at boundaries of the anchor boxes; and ii) this can be rectified by improving the way of choosing positive samples from candidate anchor boxes when training the detectors.
AB - The employment of convolutional neural networks has led to significant performance improvement on the task of object detection. However, when applying existing detectors to continuous frames in a video, we often encounter momentary miss-detection of objects, that is, objects are undetected exceptionally at a few frames, although they are correctly detected at all other frames. In this paper, we analyze the mechanism of how such miss-detection occurs. For the most popular class of detectors that are based on anchor boxes, we show the followings: i) besides apparent causes such as motion blur, occlusions, background clutters, etc., the majority of remaining miss-detection can be explained by an improper behavior of the detectors at boundaries of the anchor boxes; and ii) this can be rectified by improving the way of choosing positive samples from candidate anchor boxes when training the detectors.
UR - http://www.scopus.com/inward/record.url?scp=85085531024&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85085531024&partnerID=8YFLogxK
U2 - 10.1109/WACV45572.2020.9093553
DO - 10.1109/WACV45572.2020.9093553
M3 - Conference contribution
AN - SCOPUS:85085531024
T3 - Proceedings - 2020 IEEE Winter Conference on Applications of Computer Vision, WACV 2020
SP - 1399
EP - 1407
BT - Proceedings - 2020 IEEE Winter Conference on Applications of Computer Vision, WACV 2020
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2020 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2020
Y2 - 1 March 2020 through 5 March 2020
ER -