Feed
Azure OCR considers many factors in order to recognize text with high accuracy, such as the font size, style, noise, and orientation of the text in an image. You can try using different images and see if the text ordering is the same or different.
Unfortunately, I cannot provide a Jupyter notebook with code as this task is confidential. However, I have provided all the information you might need in the blog. Please let me know if you still require assistance.
You can find me here: rupeshgelal.com.np
曹俊杰
NLP
One question about Azure OCR, why the text order is not the common order(top-down, left-right), but billing information first and shipping information second in form region, table using column order instead row order?
Other work is really great, could you give a Jupiter notebook or a demo?