In the field of video production and image processing, video flicker removal technology has always been the key to improving video quality. With the development of technology, a new flicker removal algorithm called BlazeBVD has emerged, which can not only quickly remove flicker in videos, but also maintain the integrity of video content and the authenticity of colors.
BlazeBVD, short for Blaze Blind Video Deflickering, is an automated video deflickering technology that can effectively improve the temporal consistency of videos under unknown video degradation conditions. The emergence of this technology has brought revolutionary changes to video post-production.
Video flickering is usually caused by the shooting environment, camera hardware limitations, or image processing technology. These problems cause brightness or color inconsistencies between video frames, seriously affecting the viewing experience. Therefore, it is particularly important to develop a technology that can eliminate flickering and maintain the integrity of video content.
How BlazeBVD works
The core of BlazeBVD technology is to use the classic Scale-Time Equalization (STE) method to capture flicker and local exposure changes by processing the histogram of the video frame. This process includes the following key steps:
Histogram-assisted: BlazeBVD first compresses pixel values into illumination histograms, which are able to accurately capture flicker and local exposure variations.
STE filtering: The illumination histogram is smoothed by the STE method to generate a single frame set, filtered illumination map and exposure map.
Global and local flicker removal modules: BlazeBVD utilizes a 2D network to restore textures affected by lighting variations or local exposure issues, and uses a lightweight 3D network to correct minor temporal inconsistencies.
Adaptive Temporal Consistency: Finally, BlazeBVD processes all frames through a lightweight spatiotemporal network and improves the video coherence by designing an adaptive mask-weighted deformation loss.
Through comprehensive experiments on synthetic, real-world, and generated videos, BlazeBVD demonstrates superior qualitative and quantitative results. It not only removes flicker while maintaining video quality, but also achieves 10 times the inference speed of the state-of-the-art.
BlazeBVD technology has broad application prospects. It can be applied to video editing, film post-production, video content generation and other fields. With the continuous advancement of technology, BlazeBVD is expected to become a new standard in the field of video processing.
Paper address: https://arxiv.org/html/2403.06243v1