Classification of Leaf Hydration state based on Leaf Colour

Hi I need assistance and guidance to start my project.

Basically i need to classify images based on leaf colour.

I have 3 classes.

1. Hydrated Leaf
2. Dehydrated Leaf
3. Dry Leaf

For each classes I have 10 reference images as perceived by human eyes. From my reading I find out i need to use Colour Lab space to get Delta E value.

Please guide me how to set classifier from the reference image.

Thank you

Answer 1

[Mahesh Taparia]

Hi

You are having very less data to use deep learning, you can use those data at first. Use data augmentation to increase the dataset (atleast 40-50 image per class). You can refer to this documentation for image classification.

Try different architecture for classification, refer to this documentation for details of pretrained network. You can create your custom network to train the dataset, for this refer to this documentation.

What is delta E value? You can use 'lab' color space by using 'rgb2lab' command and use those image in the network. For more details you can refer to this documentation of 'rgb2lab'.