Dynamic hand gesture recognition for human-robot and inter-robot communication

This paper discusses inter-robot and human-robot communication by bare hand dynamic gestures. We use a Bag-of-Features and a local part model approach for bare hand dynamic hand gesture recognition from video. We used dense sampling to extract local 3D multiscale whole-part features. We adopted three dimensional histograms of a gradient orientation (3D HOG) descriptor to represent features. The K-means++ method was applied to cluster the visual words. Dynamic hand gesture classification was completed by using a Bag-of-features (BOF) and non-linear support vector machine (SVM) method. A BOF does not track the order of events. To counter the unordered events of the BOF approach, we used a multiscale local part model to preserve temporal context. Initial experimental results on the newly collected complex dataset show a higher level of recognition. We used the same above mentioned approach for inter-robot communication by using two sample hand models.