TY - GEN
T1 - Tracker evaluation for small object tracking
AU - Liu, Chang
AU - Liu, Chunlei
AU - Yang, Linlin
AU - Zhang, Baochang
N1 - Publisher Copyright:
© Springer Nature Switzerland AG 2021.
PY - 2021
Y1 - 2021
N2 - The small object problem becomes an increasingly important task because of its wide application. There are three significant challenges for small objects: 1) small objects have extremely vague and variable appearances, 2) due to the low resolution of the input images, their characteristic expression information is inadequate and, therefore, is prone to be absent after downsampling and 3) they draft drastically in the images when lens shake violently. Even though small object detection has been extensively studied, small object tracking is still in its infancy. To further explore small object tracking, we evaluate six latest trackers on OTB100 (normal object dataset) and small90 (small object dataset). According to our observation, we draw three instructive conclusions for the follow-up research of small object tracking. Firstly, due to the weak characteristics of small objects, existing trackers perform worse on small objects than on normal objects. Secondly, based on the results of ATOM, SPSTracker, DIMP, SiamFC and SiamMask, the trackers’ performance on small objects is positively correlated with that on normal objects. Thirdly, trackers tend to perform better on small object datasets when they can handle drift, occlusion and out-of-view.
AB - The small object problem becomes an increasingly important task because of its wide application. There are three significant challenges for small objects: 1) small objects have extremely vague and variable appearances, 2) due to the low resolution of the input images, their characteristic expression information is inadequate and, therefore, is prone to be absent after downsampling and 3) they draft drastically in the images when lens shake violently. Even though small object detection has been extensively studied, small object tracking is still in its infancy. To further explore small object tracking, we evaluate six latest trackers on OTB100 (normal object dataset) and small90 (small object dataset). According to our observation, we draw three instructive conclusions for the follow-up research of small object tracking. Firstly, due to the weak characteristics of small objects, existing trackers perform worse on small objects than on normal objects. Secondly, based on the results of ATOM, SPSTracker, DIMP, SiamFC and SiamMask, the trackers’ performance on small objects is positively correlated with that on normal objects. Thirdly, trackers tend to perform better on small object datasets when they can handle drift, occlusion and out-of-view.
KW - Evaluation
KW - Feature
KW - Small object
KW - Tracking
UR - https://www.scopus.com/pages/publications/85110581575
U2 - 10.1007/978-3-030-68790-8_48
DO - 10.1007/978-3-030-68790-8_48
M3 - 会议稿件
AN - SCOPUS:85110581575
SN - 9783030687892
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 622
EP - 629
BT - Pattern Recognition - ICPR International Workshops and Challenges, Proceedings
A2 - Del Bimbo, Alberto
A2 - Bertini, Marco
A2 - Sclaroff, Stan
A2 - Mei, Tao
A2 - Escalante, Hugo Jair
A2 - Cucchiara, Rita
A2 - Vezzani, Roberto
A2 - Farinella, Giovanni Maria
PB - Springer Science and Business Media Deutschland GmbH
T2 - 25th International Conference on Pattern Recognition Workshops, ICPR 2020
Y2 - 10 January 2021 through 15 January 2021
ER -