<![CDATA[The manuscript should contain an abstract. The abstract should be self-contained and citation-free and should not exceed 200 words. The abstract should state the purpose, approach, results and conclusions of the work. The author should assume that the reader has some knowledge of the subject but has not read the paper. Thus, the abstract should be intelligible and complete in it-self (no numerical references); it should not cite figures, tables, or sections of the paper. The abstract should be written using third person instead of first person. Communication barriers between Deaf individuals and people unfamiliar with Indonesian Sign Language (BISINDO) reduce the effectiveness of daily interactions. This study develops a real-time BISINDO recognition system for Android by integrating YOLOv8, MediaPipe Hands, and a one-dimensional convolutional neural network (Conv1D). YOLOv8 is used to detect the user as the region of interest, while MediaPipe Hands extracts 21 hand landmarks that are processed by the Conv1D model to classify 15 BISINDO gesture classes. A balanced dataset consisting of 900 gesture samples was used for training and evaluation. The proposed architecture contains 59,680 trainable parameters, making it lightweight and suitable for mobile deployment. Experimental results show stable convergence during training, increasing training and validation accuracy, and decreasing loss values. The classification process correctly recognizes most gesture classes, with only limited confusion among visually similar gestures. The developed Android application performs real-time gesture recognition using a smartphone camera and converts recognized gestures into text. The proposed approach demonstrates that the integration of YOLOv8, MediaPipe Hands, and Conv1D provides an efficient and practical solution for real-time BISINDO recognition and has the potential to improve communication accessibility through mobile assistive technology.]]>
Copyrights © 2026
