Python programming language is used to design a user interface for controlling the computer pointer using hand detection and classification of its gesture.

A deep learning project that classifies seven types of skin lesions using the HAM10000 dataset. Among four tested models, Swin Transformer achieved the best accuracy of 88.9%, showing AI’s potential in early skin cancer detection.

Multi-class classification on Stanford Cars Dataset

Predicts product prices using TF-IDF text features + EfficientNet image embeddings, fused into a LightGBM regressor, optimized for SMAPE and scalable e-commerce datasets.

This project leverages deep learning models, particularly EfficientNet-B0, to classify 38 plant diseases with an accuracy of 99.86%. Using advanced techniques like image segmentation (DeePLabV3+) and explainable AI (SHAP), the system aids in early disease detection, improving agricultural productivity and food security.

DeepShield is a real-time deepfake face detection system built using EfficientNet-B0 and transfer learning, achieving 92% F1-score on 140K face dataset.