The field of low-light image and video enhancement is moving towards more robust and adaptive approaches. Researchers are exploring new frameworks and techniques to improve the performance of existing methods in challenging low-light environments. One of the key directions is the development of biomimetic-inspired frameworks that can interact with action classification to achieve accurate dark video human action recognition. Another important area of research is the design of novel brightness-adaptive enhancement frameworks that can tackle local exposure inconsistencies in real-world low-light images. Semi-supervised learning is also being utilized to improve image signal processor and image enhancement tasks with minimal training data. Additionally, recursive enhancement frameworks are being proposed to decompose low-light image enhancement into a recursive task, allowing for more effective enhancement of images with a wide dynamic range. Noteworthy papers include:

• OwlSight, which proposes a robust illumination adaptation framework for dark video human action recognition, achieving state-of-the-art performance across four low-light action recognition benchmarks.
• Adaptive Low Light Enhancement via Joint Global-Local Illumination Adjustment, which introduces a novel brightness-adaptive enhancement framework that effectively models global illumination and guides local contrast enhancement.
• SemiISP/SemiIE, which realizes semi-supervised learning for image signal processor and image enhancement tasks leveraging a RAW image reconstruction method.
• Brightness Perceiving for Recursive Low-Light Image Enhancement, which proposes a brightness-perceiving-based recursive enhancement framework for high dynamic range low-light image enhancement, achieving new state-of-the-art performance on six reference and no reference metrics.