Designed and implemented a hybrid architecture combining ViT-based feature extraction with GraphSAGE classification to leverage both image-level representations and graph-structured relationships. Conducted experiments on preprocessed MRI datasets, applie

mashiat tabassum khanJagannath University

Project

Synthetic Brain Image Generation

Implemented a GAN pipeline to synthesize MRI slices for data augmentation: dataset curation, generator/discriminator implementation, training experiments and reproducible notebooks.

mashiat tabassum khanJagannath University

Project

Advanced Deep Learning Architectures

Implemented CNNs, GNNs, GANs and LSTMs for classification and generative tasks.

mashiat tabassum khanJagannath University

Project

Plant Leaf Disease Prediction

Developed CNN classification pipelines and a Streamlit demo for inference; repository includes training notebooks and deployment scripts.

CNN
Python
Streamlit