Hi everyone, I’m curious if there exists any AI model capable of converting video content into a text script?
example: First, upload a tutorial video. Next, create a script by observing and transcribing the keyboard and mouse actions shown in the video.
gpt-4-vision-preview is the OpenAI model with computer vision.
One can provide a sequence of images to one call, or a composite of thumbnails of adequate resolution, and have the AI produce some descriptive understanding of what it sees.
Eavesdropping on keypresses may be way beyond the capabilities of its image entity extraction and identification: simply producing an optical character recognition of 101 entities would be a challenge currently.
It does not accept video. I’m not immediately aware of AI that could perform this task, especially to the temporal specificity of an accompanying script.
1 Like