Identification method for pictures and based on skull image superimposition principle

Abstract

The invention discloses an identification method for pictures and based on a skull image superimposition principle. The identification method includes the steps: S1, marking and drawing measuring reference, namely marking mark points and drawing mark lines; S2, extracting facial features, namely selecting indicators, measuring length, selecting reference indicators, and extracting feature vectors; S3, outputting an identification result, namely acquiring similarity between to-be-detected picture features and criminal picture features, and outputting the identification result according to similarity. By the identification method, requirements on professional skills of measuring personnel are lowered, and measuring steps are simplified; the identification method is suitable for assisting common public security officers in quick identification. In addition, the identification method overcomes influence, on identification result, of illumination, shielding and expression, avoids complex operation by extracting different posture picture features of a same person in mode recognition methods and is simple and practical.

Description

technical field [0001] The invention belongs to the identification field, and in particular relates to a photo-identification method based on the superposition principle of cranial images. Background technique [0002] The identity identification of photos is the process of identifying whether the specified portraits appearing in two photos are the same person. In the public security business, the comparison of photos collected by suspicious persons on the spot with the files of criminals fleeing abroad belongs to the same identification technology category. [0003] Existing identification methods for photographs and photographs mainly rely on morphometric theory, such as the photographs and photographs proposed by Lan Yuwen in his monograph "Research and Application of Skull Image Superposition" (Qunyan Publishing House, July 1999). the same identification method. The method is characterized by the shape of the skull, which is not easily affected by age changes, disguise...

AI Technical Summary

Problems solved by technology

However, the database building process and measurement steps of this method are cumbersome and require specific identification equipment, so it is difficult to apply and promote in the process of searching and hunting suspects involved

Since the images to be identified may have been taken in different periods, with different shooting equipment, in different shooting environments, and in different postures of the portraits, although some scholars have tried to use pattern recognition methods to identify the same photo from photo to photo, such methods have not yet been identified. There is still a large gap between the accuracy and the business requirements of the public security department

Images