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

Plant Leaf Disease Prediction

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

mashiat tabassum khanJagannath University

Project

Advanced Deep Learning Architectures

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

Python
PyTorch
TensorFlow