On September 19, Kuaishou announced a major global upgrade to its Keling AI model, launching version 1.5. The new version features significant improvements in several areas, including video quality, dynamic effects, aesthetic performance, motion realism, and semantic understanding.

Facing the direct impact of video quality on viewing experience, especially regarding the presentation of facial details, current methods such as simply applying super-resolution networks to face datasets or independently processing each frame often struggle to balance the detail of facial reconstruction with temporal consistency. To address this challenge, the research team at Nanyang Technological University has launched the innovative framework KEEP, which utilizes the principles of Kalman filtering to achieve a 'memory' capability in facial restoration, significantly enhancing the consistency and continuity of facial details during the restoration process. The KEEP framework consists of four key modules: Encoder, Decoder, Kalman.

NVIDIA recently announced the launch of a software called Broadcast, designed to automatically eliminate background noise and room echo through AI technology. It also offers features such as background removal, replacement, and blurring, allowing users to customize their backgrounds without the need for an actual green screen or complex lighting setups. AI technology enhances audio and video quality, ensuring clear audio output through the automatic elimination of background noise and room echo.