Image electronic evidence screening method based on deep learning

Abstract

The invention provides an image electronic evidence screening method based on deep learning. The method comprises the following steps: obtaining data from electronic equipment as an evidence source, copying the evidence source, and carrying out image preprocessing to obtain an image evidence data set; transmitting the image evidence data set to a deep learning classification network for classification, and storing a classification result and classification features; meanwhile, transmitting the image evidence data set to a deep learning detection network for detection, and storing a detection result and detection features; performing threshold discrimination on the classification result and the detection result, and screening out the same target category; carrying out feature comparison onthe classification features and the detection features, and screening out the same target category; and taking the images of the same target category after threshold screening and feature comparison as electronic evidences. In order to improve the precision of image electronic evidences, two deep learning methods are adopted to screen the image electronic evidences; and in order to improve the effectiveness of the image electronic evidences, threshold screening and feature comparison methods are used for screening the image electronic evidences.

Description

technical field [0001] The present invention relates to the field of image processing technology and deep learning algorithm, in particular, relates to a method for screening electronic image evidence based on deep learning. Background technique [0002] There are often a large number of suspicious images, voices, texts, videos and other files stored in electronic devices involved in criminal cases, among which the screening and extraction of suspicious images is the most common compared to other files. With the development of 5G communication technology and the continuous improvement of network speed, the quality of chips, processors and cameras in smart devices is constantly improving, and the storage capacity is also constantly increasing. It is expected that there will be a corresponding explosion of image and video data, which will bring new challenges to the collection and screening of electronic data in electronic forensics. [0003] At present, judicial personnel at...

AI Technical Summary

Problems solved by technology

[0003] At present, judicial personnel attach great importance to the acquisition of effective electronic image evidence in the process of handling cases. However, due to the huge amount of electronic image data, traditionally, judicial personnel still adopt manual methods when extracting electronic image evidence. After copying and saving the electronic data of the image, after the electronic data of the image is obtained, the human eye is constantly inspected and compared with the set value of the valuable image electronic evidence to distinguish. This method is very simple, but it is easy to be affected by human subjectivity and emotion. Influenced by other factors, the deviation is often large and requires a large amount of labor

This artificial feature extraction method has several disadvantages: (1) The workload is mainly performed manually, which consumes a lot of personnel, has low precision and low work efficiency, resulting in a large amount of limited evidence collection time

(2) Feature extraction is only applicable to specific items in specific scenes. Once the items are changed, the algorithm also needs to be replaced, which lacks versatility

(3) Manual feature extraction requires the experience of professionals, and requires high professionalism for algorithm personnel

Images