Getting segmentation mask and class name predictions for ADE20k inference

[jin-zhe]

Hi I'm running inference based on the model trained on ADE20K dataset and I would like to:
(1) Derive the segmentation mask. i.e. matrix with same rows x cols as image but each cell reflected the class id prediction for the pixel
(2) Get the corresponding ADE20K class name given a class id

[jin-zhe]

I found my answer here: https://github.com/CSAILVision/sceneparsing

[shadydiaa]

but this coded in maltab can you share the code in python to get mask and class name , Thanks
@jin-zhe

[jin-zhe]

Hey @shadydiaa

• To get mask, at line 111 of inference.py, simply do seg_map = sess.run(raw_output_up)
• For class names, you just have to look at objectInfo150 in the repo which comes in .txt, .csv, and .mat (any one will do). There's no need for matlab