Section: New Results

Continuous action recognition

Continuous action recognition is more challenging than isolated recognition because classification and segmentation must be simultaneously carried out. We build on the well known dynamic time warping (DTW) framework and devise a novel video alignment technique, dynamic frame warping (DFW), which performs isolated recognition based on a per-frame representation of videos and on aligning a test sequence with a model sequence. Next we devise two extensions which are able to perform action recognition and video segmentation in a concomitant manner, namely one-pass DFW and two-pass DFW. Both these algorithms have their roots in the continuous speech recognition domain but, to the best of our knowledge, their extension to visual recognition of actions and activities has been overlooked. We test and illustrate the proposed methods with several public-domain datasets and we compare both the isolated and continuous recognition algorithms with several recently published methods. One journal paper was submitted in 2013 and currently is under review.