Skip to content

Afridi465/Bark-Texture-Classification

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 

Repository files navigation

Bark-Texture-Classification

Multiclass image classification of Bark-50 texture data from https://www.kaggle.com/datasets/saurabhshahane/barkvn50

Handling data imabalance

Created more images for classes which had less number of images than the average number images of all classes uisng image augmentaion. Reference for the augmentation technique: https://github.com/mdbloice/Augmentor/blob/master/notebooks/Per_Class_Augmentation_Strategy.ipynb

Transfer Learning

Loaded the pre-trained Resnet101V2 (without the top level feature layes) and added GlobalPooling2D, Dropout and final dense layer to complete the model architecture.

Outcome

Average accuracy on the training set: 98%

Average accuracy on the validation set: 90%