Facial Mole Detection Approach for Suspect Face Identification using ResNeXt-50
Anand Singh Jalal, Dilip Kumar Sharma, Bilal Sikander · 2021
The most useful applications of face recognition are criminal face identification. There is numerous facial feature help in recognizing the criminal face like nose shape, eyes color, skin tone, etc. Face moles also play a vital role in recognizing the criminal face. Previous methods were able to detect facial spots which include moles, scars, pimples, etc but fail to discriminate the face mole from other facial spots. This paper presents a mole detection classifier that isolates the mole from other facial spots. In the proposed approach, first, a person's face is identified, following which the spot on the face is determined using Normalized Cross-Correlation, and the kind of spot is determined using the Modified ResNeXt-50 CNN model. The effectiveness of the proposed model is demonstrated on the YALE image dataset. Experimental evaluation and comparison with the state-of-the-art method shows that the proposed model outperforms it.