Application and Feasibility Assessment of Artificial Intelligence in Forensic Identification and Document Examination

This paper centers on the difficulties of document examination in judicial appraisal, and proposes a composite analysis method based on artificial intelligence. Using Hough line transform for image feature extraction, the minimum error method is selected for image segmentation. Optimization of Chinese print detection is carried out, and an improved stroke width algorithm is designed based on the SWT algorithm. Experiments of Hough line transform processing of images are carried out through MATLAB to analyze the feasibility of using Hough line transform for the preprocessing of physical evidence images. The three-dimensional information of handwriting is extracted and analyzed to explore the ability of the proposed algorithm in solving the feature distortion problem of the interaction between the writing instrument and the carrier. Combining LBP and two-factor variance modeling, the impact of the minimum error method on document examination is explored. The results show that the average limiting compression ratio after Hough line transform processing is 2.9137, which is nearly improved by 70.32% and 21.86% compared with the average limiting compression ratio of wavelet transform and Fourier transform. The stroke width algorithm has the ability to compensate for material differences, and the suppression efficiency of writing fluctuation is more than 85%. After utilizing the minimum error method for character region segmentation, the 3rd to 9th dimensional features in the LBP features are only affected by the printer factor and not by the character factor. This study provides an operable path for document examination using artificial intelligence, with both theoretical innovation and practical value.