Recently, a research team led by Prof. WU Yanjun in the Intelligent Software Research Center, Institute of Software Chinese Academy of Sciences, has made a new progress in efficient hand detection. This research provides an effective way to realize accurate and real-time processing of human hand detection in video.

Hand detection is applied in many tasks such as virtual reality, human-computer interaction, and driving monitoring, to name a few. However, it is still challenging due to many difficulties such as the low-resolution, clutter background, occlusions, the varying sizes and shapes of hands due to different view angle, and inconsistent appearances due to changing illuminations.

 In the very recent work, rotation and derotation layers are added to the network to handle rotated hands. One problem with the existing deep learning based hand detection methods is that merging all the multi-scale features rudely and the complex network structure to handle rotation usually lead to high computational cost, which limits the practicality of these methods in applications that require fast processing time.

Dr. ZHANG Libo, an assistant researcher from Intelligent Software Research Center, Institute of Software Chinese Academy of Sciences present an efficient Scale Invariant Fully Convolutional Network (SIFCN) for hand detection. Dr. ZHANG and his co-authors proposed Complementary Weighted Fusion (CWF) block can make full use of the distinctive features of different scales to achieve scale invariance effectively. Specifically, the multi-scale features are merged iteratively rather than concatenated simultaneously to reduce computation overhead. Moreover, the multi-scale loss scheme is employed to accelerate the training procedure significantly. Experimental results on the VIVA and Oxford datasets show comparable performance of our method compared with the state-of-the-art methods with much higher speed.

This study was supported by the National Natural Science Foundation of China.

The work entitled “Scale Invariant Fully Convolutional Network: Detecting Hands Efficiently” was accepted in a top AI conference named AAAI Conference on Artificial Intelligence (AAAI 2019).