Enhancement of Criminal Facial Image Using Multistage Progressive V-Net for Facial Recognition by Pixel Restoration

Abstract

INTRODUCTION: Criminal activity is expanding exponentially in modern society, which leads towards a great concern about security issues. Facial recognition technology (FRT) is a powerful computer-based system that increasingly being used for recognize and match faces to solve crimes and investigations. OBJECTIVES: Due to poor image clarity and noisy pixels, the detection of criminal faces tends to be inaccurate. Hence, image enhancement techniques are required to recognize criminals with better accuracy. In the proposed model, a multistage progressive V-net based image quality enhancing technique is employed to improve accuracy. METHODS: The Convolutional Neural Network (CNN) for restoring images called MPRV-Net has three stages for a difficult balance between spatial data and highly contextualized information for image restoration tasks while recovering images. RESULTS: For image restoration tasks, including denoising, deblurring, and deraining, MPRV-Net has provided considerable performance benefits on a number of datasets. The suggested network is significant as it eliminates all three types of deviations using a single architecture. The proposed model's performance is tested using performance metrics such as accuracy, precision, recall, and specificity, obtaining 94%, 96%, 93%, and 95%. CONCLUSION: Thus, the proposed Multistage Progressive V-Net model for effectively improves the criminal Facial image for detecting criminals in public places with greater accuracy.