Abstract
In sign language, when signed letters are continuously spelled based on backhand view, a previous signed letter influences the trajectory of hand and fingers approaching the pause duration for signing the current signed letter. Since those varied trajectories are regarded as parts of the current signed letter, hand gesture during pause duration of the current signed letter is regarded as insufficient for recognition of the current signed letter. The previous signed letters, and trajectories of hand and fingers between the previous and the current signed letters should be included as data for classification. This paper proposes a method of backhand-view-based continuous-signed-letter recognition using a rewound video sequence with previous signed letter. In the method, a hand shape of previous signed letter and trajectories of finger joints moving from the previous signed letter to the current one are detected, features are then extracted, and finally, the features are classified for signed letter recognition. To evaluate performance of the proposed method, experiments with 10 participants were performed 20 times each, and the results revealed 96.07% accuracy approximately which were improved significantly from the conventional methods using forehand and backhand.
Highlights
A sign language, which is basically expressed by finger and hand gestures, consists of signed words and signed letters
This paper basically employs a series of quantized finger joint positions as input data, feed them to Long Short-Term Memory (LSTM) networks [46] for training in advance, and classify them based on the trained database into signed letters
This paper proposed a continuous signed-letter recognition method based on backhand view using rewound video sequence and previous signed letter information
Summary
A sign language, which is basically expressed by finger and hand gestures, consists of signed words and signed letters. In order to classify a signed letter in a video sequence of continuous signed letters based on the backhand view as our research problem, the authors of this paper considered to utilize the information of the current letter, and previous letter, which influences the transformation to the current letter. A. OVERALL PROPOSED METHOD As overall of the proposed method is implemented based on the mentioned basic concept, as shown, the system starts by inputting a sequence of video frames representing continuous signed letters. FEATURE EXTRACTION To extract feature for classification in the process, effective features should be determined and fed to the classifier
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
