CNN feature or handcrafted feature in DCF object tracker?
In the DCF framework, a crucial component except DCF trackers is the adopted feature map, especially when we need the tracker to be applied in practical. The powerful feature map from CNN achieves an outstanding performance when compared to the handcrafted feature (HOG, Color-Name), however, it make...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2020-08-01
|
| Series: | Engineering Science and Technology, an International Journal |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2215098619310195 |
| id |
doaj-75e79f453afa4420ab344975a2ecf379 |
|---|---|
| record_format |
Article |
| spelling |
doaj-75e79f453afa4420ab344975a2ecf3792020-11-25T02:54:53ZengElsevierEngineering Science and Technology, an International Journal2215-09862020-08-01234960965CNN feature or handcrafted feature in DCF object tracker?Yan Zhou0Hongwei Guo1Dongli Wang2Corresponding author.; College of Automation & Electronical Information, Xiangtan University, Xiangtan 411105, ChinaCollege of Automation & Electronical Information, Xiangtan University, Xiangtan 411105, ChinaCollege of Automation & Electronical Information, Xiangtan University, Xiangtan 411105, ChinaIn the DCF framework, a crucial component except DCF trackers is the adopted feature map, especially when we need the tracker to be applied in practical. The powerful feature map from CNN achieves an outstanding performance when compared to the handcrafted feature (HOG, Color-Name), however, it makes the tracker slower and could not meet the need in the real-time scene. In this paper, we visualize the respective feature maps, filter, and location score from CNN, HOG and ColorName, to make a comparison. We also dig into the detail of the target label, for balance the accuracy and robustness of the tracker. Experiments on OTB-2015 show the performance loss from handcrafted feature is acceptable for the real-time application.http://www.sciencedirect.com/science/article/pii/S2215098619310195Object trackingDCFCNNHOGRobustness |
| collection |
DOAJ |
| language |
English |
| format |
Article |
| sources |
DOAJ |
| author |
Yan Zhou Hongwei Guo Dongli Wang |
| spellingShingle |
Yan Zhou Hongwei Guo Dongli Wang CNN feature or handcrafted feature in DCF object tracker? Engineering Science and Technology, an International Journal Object tracking DCF CNN HOG Robustness |
| author_facet |
Yan Zhou Hongwei Guo Dongli Wang |
| author_sort |
Yan Zhou |
| title |
CNN feature or handcrafted feature in DCF object tracker? |
| title_short |
CNN feature or handcrafted feature in DCF object tracker? |
| title_full |
CNN feature or handcrafted feature in DCF object tracker? |
| title_fullStr |
CNN feature or handcrafted feature in DCF object tracker? |
| title_full_unstemmed |
CNN feature or handcrafted feature in DCF object tracker? |
| title_sort |
cnn feature or handcrafted feature in dcf object tracker? |
| publisher |
Elsevier |
| series |
Engineering Science and Technology, an International Journal |
| issn |
2215-0986 |
| publishDate |
2020-08-01 |
| description |
In the DCF framework, a crucial component except DCF trackers is the adopted feature map, especially when we need the tracker to be applied in practical. The powerful feature map from CNN achieves an outstanding performance when compared to the handcrafted feature (HOG, Color-Name), however, it makes the tracker slower and could not meet the need in the real-time scene. In this paper, we visualize the respective feature maps, filter, and location score from CNN, HOG and ColorName, to make a comparison. We also dig into the detail of the target label, for balance the accuracy and robustness of the tracker. Experiments on OTB-2015 show the performance loss from handcrafted feature is acceptable for the real-time application. |
| topic |
Object tracking DCF CNN HOG Robustness |
| url |
http://www.sciencedirect.com/science/article/pii/S2215098619310195 |
| work_keys_str_mv |
AT yanzhou cnnfeatureorhandcraftedfeatureindcfobjecttracker AT hongweiguo cnnfeatureorhandcraftedfeatureindcfobjecttracker AT dongliwang cnnfeatureorhandcraftedfeatureindcfobjecttracker |
| _version_ |
1724719278141210624 |