The results vary depending on the resolution of the image.
#25
by
illiliiiiil
- opened
When I cropped only the image region containing the table from a specific page and used it as input data, the model did not interpret the table well.
However, when I used the entire page as the input instead of cropping only the table, the table was recognized properly.
It seems that the results vary depending on the resolution of the input data. What was the input image resolution used during training? Or is there an optimal resolution?
illiliiiiil
changed discussion title from
Does the result of the image of the image vary?
to The results vary depending on the resolution of the image.