search

vvuonghn / AI_DocumentLayoutAnalysis

AI_DocumentLayoutAnalysis
38 stars 13 forks source link

Regarding Training & Prediction #1

Open ghost opened 5 years ago

ghost commented 5 years ago

@hongocvuong1998 Thank you for your hard work.

General:

Training:

Prediction:

Waiting for your reply

vvuonghn commented 5 years ago

@hongocvuong1998 Thank you for your hard work.

General:

  • What are the required dependencies & versions to run this program? Example: (Opencv 3 or 2? Python 3 or 2.7 ?, etc..)
  • Why didn't you use Conda instead of Visual Studio? (Conda can run on multiple platforms: Windows, Linux, Mac)

Training:

  • What is required for training?
  • How to create our own xml labels?
  • Can you provide some ground-truth data to test training?
  • Can I train from scratch? or I must use the pre-trained vgg16.npy model?
  • Can your program be trained to detect Text-Lines? or does it only detect Paragraph regions?

Prediction:

  • How can I predict using the model I trained?
  • Can I extract the images of the predicted regions?

Waiting for your reply

Can you contact with me, I will explain in detail. mail: hongocvuong1998@gmail.com Thank you for care about my project