Search

We propose an end-to-end learned image data hiding framework that embeds and extracts secrets in the latent representations of a generic neural compressor. By leveraging a perceptual loss function in conjunction with our proposed message encoder and decoder, our approach simultaneously achieves high image quality and high bit accuracy. Compared to existing techniques, our framework offers superior image secrecy and competitive watermarking robustness in the compressed domain while accelerating the embedding speed by over 50 times. These results demonstrate the potential of combining data hiding techniques and neural compression and offer new insights into developing neural compression techniques and their applications., Comment: 5 pages, 5 figure; accepted by VCIP 2023

Lossy image coding standards such as JPEG and MPEG have successfully achieved high compression rates for human consumption of multimedia data. However, with the increasing prevalence of IoT devices, drones, and self-driving cars, machines rather than humans are processing a greater portion of captured visual content. Consequently, it is crucial to pursue an efficient compressed representation that caters not only to human vision but also to image processing and machine vision tasks. Drawing inspiration from the efficient coding hypothesis in biological systems and the modeling of the sensory cortex in neural science, we repurpose the compressed latent representation to prioritize semantic relevance while preserving perceptual distance. Our proposed method, Compressed Perceptual Image Patch Similarity (CPIPS), can be derived at a minimal cost from a learned neural codec and computed significantly faster than DNN-based perceptual metrics such as LPIPS and DISTS., Comment: 7 pages, 5 figures; accepted by APSIPA ASC 2023

mage steganography is the process of hiding information which can be text, image, or video inside a cover image. The advantage of steganography over cryptography is that the intended secret message does not attract attention and is thus more suitable for secret communication in a highly-surveillant environment such as civil disobedience movements. Internet memes in social media and messaging apps have become a popular culture worldwide, so this folk custom is a good application scenario for image steganography. We try to explore and adopt the steganography techniques on the Internet memes in this work. We implement and improve the HiDDeN model by changing the Conv-BN-ReLU blocks convolution layer with a multiscale autoencoder network so that the neural network learns to embed message bits in higher-level feature space. Compared to methods that convolve feature filters on the row-pixel domain, our proposed MS-Hidden network learns to hide secrets in both low-level and high-level image features. As a result, the proposed model significantly reduces the bit-error rate to empirically 0% and the required network parameters are much less than the HiDDeN model., Comment: accepted by Media Watermarking, Security, and Forensics 2022

Smartphone is the most successful consumer electronic product in today's mobile social network era. The smartphone camera quality and its image post-processing capability is the dominant factor that impacts consumer's buying decision. However, the quality evaluation of photos taken from smartphones remains a labor-intensive work and relies on professional photographers and experts. As an extension of the prior CNN-based NR-IQA approach, we propose a multi-task deep CNN model with scene type detection as an auxiliary task. With the shared model parameters in the convolution layer, the learned feature maps could become more scene-relevant and enhance the performance. The evaluation result shows improved SROCC performance compared to traditional NR-IQA methods and single task CNN-based models., Comment: Proceedings of Computer Vision & Graphic Image Processing (CVGIP), Hsinchu, Taiwan, Aug. 16-18, 2020

JPEG has been a widely used lossy image compression codec for nearly three decades. The JPEG standard allows to use customized quantization table; however, it's still a challenging problem to find an optimal quantization table within acceptable computational cost. This work tries to solve the dilemma of balancing between computational cost and image specific optimality by introducing a new concept of texture mosaic images. Instead of optimizing a single image or a collection of representative images, the simulated annealing technique is applied to texture mosaic images to search for an optimal quantization table for each texture category. We use pre-trained VGG-16 CNN model to learn those texture features and predict the new image's texture distribution, then fuse optimal texture tables to come out with an image specific optimal quantization table. On the Kodak dataset with the quality setting $Q=95$, our experiment shows a size reduction of 23.5% over the JPEG standard table with a slightly 0.35% FSIM decrease, which is visually unperceivable. The proposed JQF method achieves per image optimality for JPEG encoding with less than one second additional timing cost. The online demo is available at https://matthorn.s3.amazonaws.com/JQF/qtbl_vis.html