Automation script for image labeler apps (pixel labeling)

Hi there,

I am learning to use image labeller (pixel labeling) to get ground truth data to train the deep learning model for my project. Currently, I have 1000 images and I labelled them manually using pixel labeling (manage to label 50 images after taking long time). This is very time consuming and I found out there is automation algorithm feature that helps to speed up this overall process. I do some research but seems like Matlab documentation only explains automation using pretrained network and it applies to video. What I want to learn is implementing the simple image processing technique below to my 1000 images and it can be applied to collection of images and not VIDEO.

%Simple script to read in image and post processed it using image processing technique (HSV). I use this script as a reference to label image using pixel labeling into "Healthy" and "Disease" category.

im = imread('testleaf2.jpg');

imHSV = rgb2hsv(im);

imT = imbinarize(imHSV(:,:,1),0.18);

figure,

subplot(121), imshow(im)

subplot(122), imshowpair(im,imT,'blend')

Please help and drop me a link for the related examples.

Thanks.

[Preeti]

Hi San Kong Lee!

I also have similar task. Were you able to find solution?

Regards

Preeti