Could steps 2 and 3 be more specific

[15931012764]

Thanks for sharing the code, it is very useful for me.
Could you write step 2 and 3 with more details in Readme?(just like step 1)
I will be appreciated if it is possible!
Thanks again!

[DashankaNadeeshanDeSilva]

Thank you for your question and I am sorry for the super late response. The step 02 is for image pre-processing and dataset creation and it has 3 tasks:

1. Image splitting to create the source dataset from the original images: Images that I used for the source data sed coming from a dataset which has large images with bigger resolutions. Therefore, I had to slice them to create new images (multiple images from a single image of the original dataset), basically from 4096 x 256 to many 64 x 64 images. This has also given me the chance to get many images. However, I did not use all the images but suitable images from the original dataset. This has been done with image_slit.py script. (The codes may look not so optimized and nice as I was an amateur)

2. Resizing images for both source and target datasets: Even though I created 64x64 images there were images that do not have 64x64 resolution such as corner images and also I had to resize mages from target images (144x144 -> 64x64). Also, I removed some images from the source dataset (entire black images).

3. Renaming images to annotate the images to create an image dataset: After creating the images for both source and target datasets, I rename them in a common fashion for a dataset. There, I labelled them with classes (with and without defects).

And these Python scripts are self executable via "python image_split.py"

I hope this might help you. If not please let me know and let's see how I can help you.