This is a study project that use SVM method to create and train a simple model for sign languages recognition. The images were first preprocessed for extracting HOG features and then become the model's input.
It can detect the direction of the finger to identify some characters that require movements (J and Z)