FACE DETECTION USING YOLO V3 in KERAS

This tutorial shows the implementation of YOLOV3 algorithm for object detection in Keras. The YOLOv3 (You Only Look Once) is a state-of-the-art, real-time object detection algorithm. The published model recognizes 80 different objects in images and videos. The problem statement here was to detect faces in a given image and then apply blurring to the background keeping the foreground i.e. the detected faces intact. For this purpose, the yolov3 algorithm was used. A Keras model akin to the darknet architecture of the yolov3 was written and then the weights were loaded into the model from a pre-trained weights file, trained on human faces data(WIDER FACE Dataset). The weights were loaded into the model layers and the model was saved. Using this saved model face detection was performed and finally using Gaussian Blur the parts of the image except the detected faces were blurred. Source code: Github: https://github.com/chinmaykumar06/face-detection-yolov3-keras