KOMPARASI ALGORITMA YOU ONLY LOOK ONCE (YOLO) PADA DETEKSI VIDEO KEKERASAN FISIK

COMPARISON OF YOU ONLY LOOK ONCE (YOLO) ALGORITHM ON PHYSICAL VIOLENCE VIDEO DETECTION

Kekerasan fisik merupakan salah satu tindak kriminal yang sering terjadi di berbagai lingkungan dan dapat memberikan dampak serius bagi korban, baik secara fisik maupun mental. Salah satu kendala dalam penanganannya adalah keterlambatan dalam mendeteksi aksi kekerasan. Solusi dari permasalahan tersebut adalah dengan mengimplementasikan algoritma terbaik antara You Only Look Once (YOLO) versi 8 dan versi 9 untuk mendeteksi kekerasan fisik melalui video secara otomatis dan cepat. Dataset yang digunakan terdiri dari dua kelas, yaitu kekerasan (violence) dan non-kekerasan (non-violence), yang telah melalui proses ekstraksi, pembersihan data, dan pelabelan menggunakan Roboflow. Model dilatih menggunakan Google Collaboratory, dan hasil pelatihan dievaluasi menggunakan metrik mAP, precision, recall, dan F1-Score. Berdasarkan hasil pengujian, YOLOv9 memperoleh performa terbaik dengan precision sebesar 0.8096, recall sebesar 0.8665, F1-score sebesar 0,8363, dan mAP sebesar 0.8117. Sistem deteksi kemudian diimplementasikan ke dalam aplikasi berbasis web menggunakan framework Flask, yang memungkinkan pengguna untuk mengunggah video dan mendeteksi tindakan kekerasan secara otomatis. Hasil pengujian menunjukkan bahwa aplikasi berjalan sesuai dengan fungsinya dan mampu mendeteksi kekerasan fisik dengan baik. Penelitian ini diharapkan dapat menjadi solusi pendukung dalam sistem pengawasan keamanan berbasis video.

Kata kunci: Kekerasan Fisik, YOLOv8, YOLOv9, Deteksi Video, Deep learning, Object Detection.

Physical violence is one of the crimes that often occurs in various environments and can have a serious impact on victims, both physically and mentally. One of the obstacles in handling it is the delay in detecting acts of violence. The solution to this problem is to implement the best algorithm between You Only Look Once (YOLO) version 8 and version 9 to detect physical violence through video automatically and quickly. The dataset used consists of two classes, namely violence and non-violence, which have gone through the process of extraction, data cleaning, and labeling using Roboflow. The model was trained using Google Collaboratory, and the training results were evaluated using mAP, precision, recall, and F1-score metrics. Based on the test results, YOLOv9 obtained the best performance with a precision of 0.8096, recall of 0.8665, F1-score of 0.8363, and mAP of 0.8117. The detection system is then implemented into a web-based application using the Flask framework, which allows users to Upload videos and detect acts of violence automatically. The test results show that the application runs according to its function and is able to detect physical violence well. This research is expected to be a supporting solution in video-based security surveillance systems.

Keywords: Physical Violence, YOLOv8, YOLOv9, Video Detection, Deep learning, Object Detection.