機械知覚&ロボティクスグループ
中部大学

Deep Learning 学術論文(E)

Object-ABN: Learning to Generate Sharp Attention Maps for Action Recognition

Author
Tomoya Nitta, Tsubasa Hirakawa, Hironobu Fujiyoshi, Toru Tamaki
Publication
IEICE TRANSACTIONS on Information and Systems, Vol. E106-D, No. 3, pp. 391-400, 2023

Download: PDF (English)

In this paper we propose an extension of the Attention Branch Network (ABN) by using instance segmentation for generating sharper attention maps for action recognition. Methods for visual explanation such as Grad-CAM usually generate blurry maps which are not intuitive for humans to understand, particularly in recognizing actions of people in videos. Our proposed method, Object-ABN, tackles this issue by introducing a new mask loss that makes the generated attention maps close to the instance segmentation result. Further the Prototype Conformity (PC) loss and multiple attention maps are introduced to enhance the sharpness of the maps and improve the performance of classification. Experimental results with UCF101 and SSv2 shows that the generated maps by the proposed method are much clearer qualitatively and quantitatively than those of the original ABN.

前の研究 次の研究