July 19, 2018
July 18, 2018
Day #285 - Experimenting on Custom Segmentation using Unet
1. Data Annotation - Leveraged tool - eraseimage by Akshay Bhat (Nice Effort)
Segment and Save the Output Image
Step 2 - Labelling in Sequence (Follow the same structure replace with Image / Label) - Labelling Training from 0 to 28
3. Labelling Test from 0 to 4
Performed a rename as all files were in jpeg, used ren *.* to *.png to port to png format
4. Performed Data Augmentation and took those masked images and labels (Replaced Earlier ones)
Output Results
Happy Learning!!!
Pick Green - Foreground, Red - Background to mark regions
Segment and Save the Output Image
Step 2 - Labelling in Sequence (Follow the same structure replace with Image / Label) - Labelling Training from 0 to 28
Performed a rename as all files were in jpeg, used ren *.* to *.png to port to png format
4. Performed Data Augmentation and took those masked images and labels (Replaced Earlier ones)
Output Results
Happy Learning!!!
Subscribe to:
Posts (Atom)
