Automatic Hand Sign Recognition: Identify Unusuality Through Latent Cognizance
Sign language is a main communication channel among a hearing disability community. Automatic sign language transcription could facilitate better communication and understanding between a hearing disability community and a hearing majority.
As a recent work in automatic sign language transcription has discussed, effectively handling or identifying a non-sign posture is one of the key issues. A non-sign posture is a posture unintended for sign reading and does not belong to any valid sign. A non-sign posture may arise during a sign transition or simply from an unaware posture. Confidence ratio (CR) has been proposed to mitigate the issue. CR is simple to compute and readily available without extra training. However, CR is reported to only partially address the problem. In addition, CR formulation is susceptible to computational instability.
This article proposes alternative formulations to CR, investigates an issue of non-sign identification for Thai Finger Spelling recognition, explores potential solutions and has found a promising direction. Not only does this finding address the issue of non-sign identification, it also provide an insight behind a well-learned inference machine, revealing hidden meaning and new interpretation of the underlying mechanism. Our proposed methods are evaluated and shown to be effective for non-sign detection.
KeywordsHand sign recognition Thai Finger Spelling Open-set detection Novelty detection Zero-shot learning Inference interpretation
- 1.Bendale, A., Boult, T.E.: Towards open set deep networks. CoRR abs/1511.06233 (2015). http://arxiv.org/abs/1511.06233
- 7.Nakjai, P., Katanyukul, T.: T. J Sign Process Syst (2018). https://doi.org/10.1007/s11265-018-1375-6
- 10.Schölkopf, B., Williamson, R., Smola, A., Shawe-Taylor, J., Platt, J.: Support vector method for novelty detection. In: Proceedings of the 12th International Conference on Neural Information Processing Systems, NIPS 1999, pp. 582–588. MIT Press (1999)Google Scholar
- 11.Silanon, K.: Thai finger-spelling recognition using a cascaded classifier based on histogram of orientation gradient features. Comput. Intell. Neurosci. 11 (2017). https://doi.org/10.1155/2017/9026375
- 12.Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recognition. CoRR abs/1409.1556 (2014). http://arxiv.org/abs/1409.1556
- 13.Xian, Y., Schiele, B., Akata, Z.: Zero-shot learning - the good, the bad and the ugly. In: IEEE Computer Vision and Pattern Recognition (CVPR) (2017)Google Scholar