Unraveling the Gestures: Indonesian Sign Language Classification via Faster Region-CNN

Communication activities are one of the needs for human being. However, differences in speech styles between non-disabled people and people with disabilities like the mute and the deaf where they use sign language to communicate will be easier if there’s tools that can help them to communicate. Using Faster R-CNN (Region Convolutional Neural Network) and utilizing technology from camera assistance which is processed using computer vision and machine learning in the sign language reading process, that can become bridge for communication activities between one another. Faster R-CNN method used for processing sign language classification, in this case SIBI (Sistem Isyarat Bahasa Indonesia) is used for dataset training. The results of this study produce a model with mAP = 0.50 IoU amounting to 100.00% and mAP = 0.75 IoU amounting to 98.07% under predetermined conditions.