Autoregressive Generative Model for Steganographic Data Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing steganalysis techniques are inadequate in effectively removing hidden data from digital content, particularly images, as they often damage the image quality and require complex detection processes.
Innovation Solution
A method utilizing an autoregressive generative model, such as PixelCNN, is employed to restore target pixels in an image based on their adjacent pixel value distribution, allowing for the removal of hidden data without significantly altering the image quality, using a computer device that inputs the source image and selects target pixels for regeneration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing steganalysis techniques are used to detect hidden data, then hidden data can be detected, but image quality is damaged and detection process becomes complex
Solution Approach 1:
The patent replaces traditional mechanical steganalysis detection methods with a generative adversarial network (GAN) based approach. The GAN learns the underlying distribution of natural images and generates realistic image patches, substituting complex detection algorithms with a learning-based system that preserves image quality while detecting steganographic modifications
Solution Approach 2:
The patent introduces an intermediary generative model that acts as a mediator between the stego image and the detection process. Instead of directly analyzing the stego image for hidden data, the GAN generates intermediate realistic image patches that serve as a reference, allowing indirect detection of steganographic alterations while preserving the original image quality
2Measurement precision
If complex steganalysis detection processes are applied, then hidden data detection accuracy improves, but processing complexity increases
Solution Approach 1:
The patent changes the fundamental parameters of the detection approach by using a generative model trained on natural image statistics rather than traditional steganalysis parameters. The GAN learns high-level image characteristics and uses these to detect deviations caused by steganography, simplifying the detection process while maintaining or improving accuracy
Solution Approach 2:
The patent creates a synthetic copy of natural image patches through the generative model. Instead of directly analyzing complex stego patterns, the system generates realistic image copies that serve as comparison references, simplifying the detection process by using intuitive image generation rather than complex mathematical analysis
Data Source
AI summary
Disclosed is a hidden data removal method based on an autoregressive generative model which is performed by a computer device. The hidden data removal method includes receiving a source image, randomly selecting a target pixel from the source image, and inputting the source image and an identifier of the target pixel to an autoregressive generative model and restoring the target pixel from the source image. The source image is an image in which steganography-based data is hidden, and the autoregressive generative model restores the target pixel on the basis of a pixel value distribution for pixels adjacent to the target pixel in the source image.