Real-time Gender Identification from Face Images using you only look once (yolo)

Varnima E.K., C. Ramachandran

20202020 4th International Conference on Trends in Electronics and Informatics (ICOEI)(48184)18 citationsDOI

Abstract

Gender identification is an important area under which the researches are still going on. There are many gender prediction systems made using different architectures. This paper presents a real-time system which can be used for gender prediction from face images. The technique used is You Only Look Once (YOLO) v3 object detection algorithm. Darknet is used for training. Keras and OpenCV are used for testing. The dataset used is a combination of IMDb, Google Indian images and some Indian custom images taken using mobile camera. The test image accuracy was found to be 84.69%. Labelimg is the software used for labeling the face images. The aim of the proposed work is to use the system in case of monitoring, security concerned areas etc.

Topics & Concepts

Computer scienceFace (sociological concept)Identification (biology)Artificial intelligenceComputer visionSoftwareFace detectionFacial recognition systemObject detectionImage (mathematics)Object (grammar)Pattern recognition (psychology)Social scienceProgramming languageBiologyBotanySociologyFace recognition and analysisFace and Expression RecognitionBiometric Identification and Security