flowerclassifcation

Flower classification using deep learning googlenet in matlab

flowerclassifcation

Flower classification using deep learning googlenet in matlab

This model is used to classify 102 categories of flowers with the acucracy of approximately ~90% for validation and training accuracy. The link to all of the files: https://www.robots.ox.ac.uk/~vgg/data/flowers/102/

The model is running on original images, but can also be applied to segmented images of flowers (will be commented out in the code). There is also two files: GradCam and Oclussion Sensitivty map files that will randomly select 10 images and will put it on a figure.

How to use:

1. Download the file with the images from the link above. Put all the files in the same folder
2. Download googlenet if you haven't done it yet
3. Name the folder with flowers "flowers102" and the segmented images "segmented".
4. start the file with name flowers102.m
5. Manually start the files names imagelabels.mat and training102flowers.mat.
6. Check that after launching these files you have lgraph_1 and lables in your Workplace.
7. Run the program and enjoy ;)