Classification of Anomalies in Gastrointestinal Tract through Endoscopic Imagery with Deep Learning

We used the human gastrointestinal (GI) tract endoscopic imagery in order to detect different anomaly types.we used KVASIR dataset for that task. In this task we used Deep Convolutional Neural Network with transfer learning

The KVASIR dataset consists of 8,000 annotated GI tract images in 8 different classes (different annomalies) where 1000 images belong to each class.

It contains anatomical landmarks:

• dyed-lifted-polyps
• normal-cecum
• normal-pylorus
• normal-z-line
• esophagitis
• polyps
• ulcerative colitis
• dyed-resection-margins

You can download the image dataset from here

Libraries:

• Keras: Keras is a popular deep learning framework.

• Numpy & Scikit-learn: For numerical processing and machine learning

• Matplotlib: The most popular plotting tool for Python.

Explanation Coming Soon 😜

There are some issue when view .ipynb 😑 ("Sorry, something went wrong. Reload?") Probably a problem with the GitHub notebook viewing tool - sometimes github fails to render the ipynb notebooks, I believe that is some temporary (and recurring) issue with their backend 😤