How To Use A Chat To Recognize Images
Basically we’ll make a database, one that holds a mapping between tasks and results.
An example task: “Find letters in image #1004248”
Result in this case: “[1 2 3 4] [100 3 5 7] …” (positions of letters in image)
Then we just collect solutions. We use an interface where a lot of programs can give us information — like the result of a segmentation algorithm.
[program upcoming]