Tetras.AI's video frame interpolation improves video frame rate without altering shooting conditions by using AI optical flow and scene recognition technology to synthesize intermediate frames. This solves the image quality problems caused by insufficient video frame rate, such as video freezing, discontinuous motion, and blurring.
Product Features:
1. Real-time, self-adaptive interpolation at multiple frame rates;
2. Industry-leading capabilities in handling small objects at high-speed motion, subtitles, background distortion, and other challenging issues;
3. Multi-scene support: Screen frame interpolation to match display technology, video content enhancement in slow motion, standard movie frame interpolation of ultra-high-definition videos, along with gaming and animation frame interpolation for a smoother experience.
Integration into multiple platforms, such as Qualcomm, MTK, Intel, AMD, etc., is supported. The product can run on mid-to-high-end platforms in real-time through multiple computing units, such as CPU/GPU/NPU, and has already been adapted for multiple video playback applications. The video frame interpolation technology can also be used as an IP to be integrated into ISP chips to further reduce power consumption and improve performance.
Intelligently recognizes and optimizes a variety of scenes, such as movies/TV shows, sporting events, games, and vlogs, based on video characteristics, and improves video frame rates in real-time.
Embedding scene recognition, post-production content processing, real-time optical flow estimation, ghost removal, and other modules; only mobile computing power is used; supports frame interpolation up to 120FPS@2k from 30FPS@2k.
Stabilizes video after frame interpolation by identifying, locating and optimizing the post-production video content. Subtitles, logos, game components, and other content are more stable and smoother.