Latent label mining for group activity recognition in basketball videos

Motion information has been widely exploited for group activity recognition in sports video. However, order to model and extract the various motion between adjacent frames, existing algorithms only use coarse video-level labels as supervision cues. This may lead ambiguity of extracted features omission changing rules patterns that are also important video recognition. In this paper, a latent label mining strategy basketball videos is proposed. The authors' novel allows them obtain set marking different frames an unsupervised way, build frame-level representations with two separate levels signal. Firstly, digged using hierarchical clustering technique. generated then taken signal train deep CNN extraction. Lastly, fed into LSTM network spatio-temporal representation Experimental results on public NCAA dataset demonstrate proposed algorithm achieves state-of-the-art performance.