An enhanced self-attention and A2J approach for 3D hand pose estimation
Published in Multimedia Tools and Applications, 2021
This paper focuses on augmenting convolution with self-attention to capture long-range dependencies in a depth image and extends anchor points to the depth dimension to regress 3D hand joint locations.
Recommended citation: M-Y Ng., C.-B. Chng, W-K Koh, CK Chui, MCH Chua. "An enhanced self-attention and A2J approach for 3D hand pose estimation." Multimedia Tools and Applications, 2021.