10 January 2022 in IEEE Transactions on Pattern Analysis and Machine Intelligence
Multiple image hiding aims to hide multiple secret images into a single cover image, and then recover all secret images perfectly. Such high-capacity hiding may easily lead to contour shadows or color distortion, which makes multiple image hiding a very challenging task. In this paper, we propose a novel multiple image hiding framework based on invertible neural network, namely DeepMIH. Specifically, we develop an invertible hiding neural network (IHNN) to innovatively model the image concealing and revealing as its forward and backward processes, making them fully coupled and reversible. The IHNN is highly flexible, which can be cascaded as many times as required to achieve the hiding of multiple images. To enhance the invisibility, we design an importance map (IM) module to guide the current image hiding based on the previous image hiding results. In addition, we find that the image hidden in the high-frequency sub-bands tends to achieve better hiding performance, and thus propose a low-frequency wavelet loss to constrain that no secret information is hidden in the low-frequency sub-bands. Experimental results show that our DeepMIH significantly outperforms other state-of-the-art methods, in terms of hiding invisibility, security and recovery accuracy on a variety of datasets.
20 September 2023 in IEEE Transactions on Circuits and Systems for Video Technology
Image compression at extremely low bit-rates has always been a challenging task in bandwidth limited scenarios, such as aerospace and deep-sea explorations. Recent years have seen great success of deep learning in image compression, however, few of them are specially designed for extremely low bit-rate conditions. To solve this issue, in this paper, we propose a novel invertible image generation based framework for extremely low bit-rate image compression. The proposed framework is composed of three modules, including an invertible image generation (IIG) module, a generated image compression (GIC) module and a compressed image adjustment (CIA) module. The role of IIG module is to generate a compression-friendly image from the original image. In the IIG module, image generation and restoration are modelled as two mutually reversible processes to avoid the information loss. After the IIG module, the GIC module is employed to compress the generated images to save the coding bit-rates. After that, the CIA module is used to shrink the quality gap between the compressed generated image and the un-compressed image. Finally, the image from the CIA module is sent back to the IIG module to restore the original image. The experimental results on three different datasets show that the proposed framework achieves state-of-the-art performance in image compression with extremely low bit-rates. We also extend the proposed framework to feature compression towards object detection, which saves 90% bit-rates than the VVC standard with the same detection accuracy.
1 November 2024 in Chinese Journal of Electronics
Nowadays, people are getting used to upload images to a third party for post-processing, such as image denoising and super-resolution. This may easily lead to the disclosure of the privacy in the confidential images. One possible solution is to encrypt the image before sending it to the third party, the encrypted image can be easily detected by a malicious attacker in the transmission channel. We propose a confidential image super-resolution method named HSR-Net which firstly hide the secret image and then super-resolve it in the hidden domain. The method is composed of three important modules: image hiding module (IHM), image super-resolution module (ISM), and image revealing module (IRM). The IHM aims to encode secret image and hide it into a cover image to generate the stego image. The stego image looks similar to the cover image but contains the information of the secret image. The third party uses the ISM to perform image super-resolution on the stego image. The user can reveal the super-resolved secret image from the stego image. The proposed HSR-Net method has two advantages. It ensures that the third party cannot directly operate on the secret image, thus protecting the user's privacy. Due to the similarity between the stego image and cover image, we can reduce the attacker's suspicion to further improve the image security. The experimental results were tested on DIV2K dataset and Flickr2K dataset. The peak signal-to-noise ratios (PSNR) of IHM, ISM, and IRM are 38.81 dB, 28.91 dB, and 23.51 dB, respectively, which verify that the proposed HSR-Net method is able to achieve image super-resolution and protect user's privacy simultaneouly.
31 March 2026 in IEEE Transactions on Pattern Analysis and Machine Intelligence
Video stereo matching is the task of estimating consistent disparity maps from rectified stereo videos. There is considerable scope for improvement in both datasets and methods within this area. Recent learning-based methods often focus on optimizing performance for independent stereo pairs, leading to temporal inconsistencies in videos. Existing video methods typically employ sliding window operation over time dimension, which can result in low-frequency oscillations corresponding to the window size. To address these challenges, we propose a bidirectional alignment mechanism for adjacent frames as a fundamental operation. Building on this, we introduce a novel video processing framework, BiDAStereo, and a plugin stabilizer network, BiDAStabilizer, compatible with general image-based methods. Regarding datasets, current synthetic object-based and indoor datasets are commonly used for training and benchmarking, with a lack of outdoor nature scenarios. To bridge this gap, we present a realistic synthetic dataset and benchmark focused on natural scenes, along with a real-world dataset captured by a stereo camera in diverse urban scenes for qualitative evaluation. Extensive experiments on in-domain, out-of-domain, and robustness evaluation demonstrate the contribution of our methods and datasets, showcasing improvements in prediction quality and achieving state-of-the-art results on various commonly used benchmarks. The project page, demos, code, and datasets are available at: https://tomtomtommi.github.io/BiDAVideo/.