Embedded Files
Skin Cancer Diagnosis from Dermoscopic Images using Deep Learning
Data:
HAM10000 dataset (Human Against Machine with 10000 training images)
7 different skin lesion classes
Source: https://www.kaggle.com/datasets/kmader/skin-cancer-mnist-ham10000/data
Tools: TensorFlow, Lime
Objectives:
Classify the images into 7 different categories using CNN architecture
Using transfer learning to improve performance
Using data augmentation for better generalization
Applying Explainable AI (XAI) tools such as Lime and SHAP for transparency
Using Generative Adversarial Networks (GAN) to produce synthetic images
Results: The CNN architecture achieves -
Validation accuracy: 74.69%
Testing accuracy: 73.19%
Face and Text Detection in Newspaper Images
Task: When a search word is entered, the program scans the newspaper pages for that word. If the word is found on a particular page, all the faces detected on that page are compiled and returned as a contact sheet.
Tools: OpenCV, pytesseract, Pillow
Synopsis:
A zip file of newspaper images is taken as input. Then they're processed using the pillow library.
pytesseract is used for optical character recognition from the images.
OpenCV is used to detect faces in the images.
A contact sheet is created using the pillow library to accumulate all the faces detected on a particular page.
Page updated
Google Sites
Report abuse