ALGORITHM FOR FACIAL IMAGE NORMALIZATION AND ANTHROPOMETRIC POINT DETECTION

Abstract

Facial image normalization and anthropometric point analysis algorithms are widely used in fields such as biometric security, medical diagnostics, digital animation, and face recognition. This article reviews the main steps of facial image normalization, including illumination stabilization, geometric transformation, and image scaling techniques. It also discusses the principles and applications of advanced facial anthropometric point detection algorithms, such as the DLib 68-point model, MediaPipe Face Mesh, Active Shape Models (ASM), and MTCNN. The article is devoted to the study of modern facial analysis methods, their advantages and limitations, and provides theoretical and practical foundations for face identification and face geometry analysis. This study indicates the directions for identifying and solving current problems in the field. Facial image normalization and anthropometric point analysis are important steps in modern computer vision and biometric systems. This article aims to study modern facial analysis techniques and covers the main stages of image normalization, including important processes such as illumination stabilization, geometric transformation, image rotation and skew correction. The theoretical foundations, operating methods and advantages of algorithms used to determine important anthropometric points on the face - for example, the DLib 68-point model, MediaPipe Face Mesh and MTCNN - are analyzed. Also, their practical applications in face identification, facial shape anomaly detection, emotion analysis and medical diagnostics are discussed in detail. The article shows Active Shape Models (ASM), Principal Component Analysis (PCA), histogram equalization and other advanced techniques as important tools for more accurate and efficient implementation of the facial analysis process. The study proposes innovative approaches aimed at identifying and solving current problems in the field of facial analysis. This article serves to create a scientific and practical basis for further improving facial recognition and analysis technologies.