Анализ применения методов стеганографии изображений в цифровых криминалистических исследованиях
Аннотация
Images are classified according to loaded and unloaded, and are divided into: uncovered-cover images and loaded-stego. In today's society, the functions of mobile phone cameras are becoming more and more abundant, and the photos sent by people will be pre-processed, which causes some steganography analyzers to make misjudgments on unencrypted images, which reduces the accuracy of steganography analysis. Therefore, front-end detection of tampered images is becoming more and more important.
This thesis combines digital forensics to study steganography, and uses the CNN model to detect untampered images for steganography analysis.
The main work includes:
1. In the research of image tampering detection, this paper performs SIFT feature extraction, feature point matching, and clustering on the image, detects and removes the tampered image, and retains the non-tampered image.
2. Taking the stego image as the input of the CNN model, using a 5 × 5 convolution kernel to perform unfilled high-pass filter preprocessing on the input image, and then sequentially passing through the convolution layer and the fully connected layer, and finally get the judgment results of the cover and stego images, respectively. Realize the steganography judgment of the image.
3 Experiments show that the method of this paper improves the accuracy of steganalysis
This thesis combines digital forensics to study steganography, and uses the CNN model to detect untampered images for steganography analysis.
The main work includes:
1. In the research of image tampering detection, this paper performs SIFT feature extraction, feature point matching, and clustering on the image, detects and removes the tampered image, and retains the non-tampered image.
2. Taking the stego image as the input of the CNN model, using a 5 × 5 convolution kernel to perform unfilled high-pass filter preprocessing on the input image, and then sequentially passing through the convolution layer and the fully connected layer, and finally get the judgment results of the cover and stego images, respectively. Realize the steganography judgment of the image.
3 Experiments show that the method of this paper improves the accuracy of steganalysis