Japanese OCR with CenterNet

An image-based document classification system that uses deep learning models (EfficientNet, ResNet, Vision Transformers) to automatically classify financial documents into predefined categories like Report, Resume, Letter, and more. Built with FastAPI for API integration and PyTorch for model training.

Recycling fails when different types of waste get mixed together. Manual sorting is slow and unhygienic, so a smart automated system is needed. RecycleVision is a deep learning based image classifier that identifies different types of waste and makes automatic segregation possible for smart bins and recycling plants.

Custom Efficient Neural Network For Histopathological Lung Cancer Classification

Flood detector using effientnet as semantic segmentation model

Brain MRI tumor classification system built with EfficientNetB0 and Explainable AI (Grad-CAM). Features medical image preprocessing (CLAHE) and deployment via Streamlit & Hugging Face Hub.

Image classifier for dog breeds using EfficientNetB3 pre-trained model and transfer learning. The model achieves 94.8% training accuracy and 90% validation accuracy across 120 dog breeds. Features automatic pipeline for data loading, augmentation, training and evaluation using TensorFlow/Keras.

This is a warehouse for EfficientDet-pytorch-model, can be used to train your object-detection-dataset

Binary classification of colon histopathology images into Colon Adenocarcinoma vs Benign Tissue using transfer learning on the LC25000 dataset.

Deep-learning galaxy morphology classifier — EfficientNet-B0 fine-tuned on Galaxy Zoo 2, served via FastAPI with a Next.js + Three.js front-end that classifies a live gallery and any image you upload.