What's the difference between Mate 20AI main photographer 2.0 and 1.0?

Compared with the 1.0 version, AI Photographer 2.0 has added more than 80 scene recognition capabilities in 8 categories. Thanks to the powerful computing power of dual-core NPU on Kirin 980 hardware platform, the AI photographer function is realized, and the image semantic segmentation at pixel level can be run in real time when shooting video. A leap from scene recognition to object recognition.

Its functions include:

1. Image local effect adjustment

2. Artificial intelligence 4D focus tracking

3. Artificial intelligence movie mode

4.AI intelligent image transmission

Note: Pixel-level semantic segmentation is different from ordinary scene recognition.

Ordinary scene recognition: only the image shooting scene and the main target categories in the scene can be recognized.

Pixel level semantic segmentation:

We can infer every pixel in the image, not only to judge the categories of various objects in the scene, but also to judge their types, positions and boundary information, which can be used in more applications.