TY - GEN
T1 - CCF-Net
T2 - 29th International Conference on MultiMedia Modeling, MMM 2023
AU - Ye, Kai
AU - Ji, Haoqin
AU - Li, Yuan
AU - Wang, Lei
AU - Liu, Peng
AU - Shen, Linlin
N1 - Publisher Copyright:
© 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2023
Y1 - 2023
N2 - Human parts detection has made remarkable progress due to the development of deep convolutional networks. However, many SOTA detection methods require large computational cost and are still difficult to be deployed to edge devices with limited computing resources. In this paper, we propose a lightweight Cascade Center-based Framework, called CCF-Net, for human parts detection. Firstly, a Gaussian-Induced penalty strategy is designed to ensure that the network can handle objects of various scales. Then, we use Cascade Attention Module to capture relations between different feature maps, which refines intermediate features. With our novel cross-dataset training strategy, our framework fully explores the datasets with incomplete annotations and achieves better performance. Furthermore, Center-based Knowledge Distillation is proposed to enable student models to learn better representation without additional cost. Experiments show that our method achieves a new SOTA performance on Human-Parts and COCO Human Parts benchmarks(The Datasets used in this paper were downloaded and experimented on by Kai Ye from Shenzhen University.).
AB - Human parts detection has made remarkable progress due to the development of deep convolutional networks. However, many SOTA detection methods require large computational cost and are still difficult to be deployed to edge devices with limited computing resources. In this paper, we propose a lightweight Cascade Center-based Framework, called CCF-Net, for human parts detection. Firstly, a Gaussian-Induced penalty strategy is designed to ensure that the network can handle objects of various scales. Then, we use Cascade Attention Module to capture relations between different feature maps, which refines intermediate features. With our novel cross-dataset training strategy, our framework fully explores the datasets with incomplete annotations and achieves better performance. Furthermore, Center-based Knowledge Distillation is proposed to enable student models to learn better representation without additional cost. Experiments show that our method achieves a new SOTA performance on Human-Parts and COCO Human Parts benchmarks(The Datasets used in this paper were downloaded and experimented on by Kai Ye from Shenzhen University.).
KW - Human parts
KW - Knowledge distillation
KW - Object detection
UR - http://www.scopus.com/inward/record.url?scp=85152539367&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-27818-1_15
DO - 10.1007/978-3-031-27818-1_15
M3 - Conference contribution
AN - SCOPUS:85152539367
SN - 9783031278174
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 177
EP - 189
BT - MultiMedia Modeling - 29th International Conference, MMM 2023, Proceedings
A2 - Dang-Nguyen, Duc-Tien
A2 - Gurrin, Cathal
A2 - Smeaton, Alan F.
A2 - Larson, Martha
A2 - Rudinac, Stevan
A2 - Dao, Minh-Son
A2 - Trattner, Christoph
A2 - Chen, Phoebe
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 9 January 2023 through 12 January 2023
ER -