Looking on the internet about photo restoration and video upscale, i found the stable diffusion engine.
i install and train a model in a week, to learn more about a female model.
after some runs, with online photos about that lady, i pick a old video and selected a frame, and start to make other frames.
in seconds, it was generates a real, upscaled and cleaning shot about the video, and comparing about the veai, it looks clearly more efficient and smarter.
i took a seconds to realizes that ai, can make a full video upscale, with face and body reconstructed and create new frames from the zero.
the script is very hard to work, and the problem starts there. so i was thinking that the veai can pickup some tools in this tech and add to the topaz word
the stable diffusion is in 1.5 version, and the open source is in github. if the lab can use ou incorporate to the video enchance ai, it will be a jump to the future
