25 December 2025 in IEEE Signal Processing Letters
As an emerging technology, Video Face Super-Resolution (VFSR) aims to reconstruct high-resolution facial images from low-quality video sequences while maintaining identity consistency, which makes it applicable to scenarios such as surveillance, video conferencing, and film restoration. Compared with image-based face restoration and general video super-resolution, VFSR is more challenging because it requires accurate facial detail reconstruction, strict identity preservation, and computational efficiency under varying poses and expressions. To address these challenges, we propose a High-Precision Identity Preserving VFSR framework (HPIP), which integrates a Multi-Scale Prediction Module (MPM) and an Identity Preservation Module (IPM). The MPM focuses on identity-critical facial regions (e.g., eyes, nose, and mouth) and leverages multi-scale feature prediction to improve reconstruction accuracy and robustness while maintaining computational efficiency. The IPM further projects features into a latent representation space, generating temporally consistent dictionary features and enhancing temporal coherence. Extensive experiments demonstrate that HPIP achieves superior performance in both qualitative and quantitative evaluations, producing visually pleasing facial details while maintaining an efficient and lightweight design.
27 October 2025 in IEEE Signal Processing Letters
Accurate face detection in low-light environments remains a challenging task due to severely degraded visibility of facial features. Conventional cascade-based methods, in which enhancement and detection are applied sequentially, are often found to achieve suboptimal accuracy. Meanwhile, detection only approaches are frequently observed to yield imprecise localization in the absence of explicit illumination adjustment. Although several joint enhancement-detection frameworks have recently been proposed, effective collaboration between the two tasks is still lacking in many cases, limiting overall efficacy. To overcome these limitations, a Value-aware Enhancement and Detection Network (VED-Net) is proposed—an end-to-end dual task architecture designed for simultaneous face enhancement and detection in low-light images. Central to the framework is a Retinex-based V-channel Enhancement and Extraction (VEE) module, through which illumination is improved and refined V-channel features are extracted in the HSV color space. The model is further equipped with complementary enhancement and detection branches that cooperatively utilize the output from the VEE module to enhance feature representation for both tasks. Experimental results demonstrate that significant improvements in both detection accuracy and image quality are achieved by VED-Net, confirming its potential for practical applications in nighttime scenarios.