🫁 Automate ETT and Carina segmentation on chest radiographs for faster, accurate assessments, improving patient care and treatment efficiency.

📊 Forecast time-series data using LSTM models in PyTorch; generate, train, and visualize predictions with key metrics for accurate insights.

Image augmentation extension for the image-dataset-converter library.

Deep Learning for Automatic Pneumonia Detection, RSNA challenge

Next-generation Albumentations: dual-licensed for open-source and commercial use

Implements a UNet-based medical image segmentation framework for precise detection of the carina and endotracheal tube tip, supporting automated clinical evaluation of airway placement.

Shuffle PatchMix (SPM) for Source-Free Domain Adaptation (ICIP 2025); patch-shuffle augmentation + confidence-margin pseudo-labels. New SOTA on PACS (+7.3%), strong results on DomainNet-126 and VisDA-C.

An MCP-compatible image augmentation tool powered by Albumentations. Built for Claude, Kiro, and other AI agents.

ML data processing (For Computer Vision)

A Python module implementing "CAP-VSTNet: Content Affinity Preserved Versatile Style Transfer" that allows for modular implementation of style transfer as an image augmentation in deep learning pipelines, with a fully Pytorch-based framework for image and video training and inference.

Fast and flexible image augmentation library. Paper about the library: https://www.mdpi.com/2078-2489/11/2/125

A high-performance image processing library designed to optimize and extend the Albumentations library with specialized functions for advanced image transformations. Perfect for developers working in computer vision who require efficient and scalable image augmentation.

INCOIS_AAIDeS (Automated Animal Identification and Detection of Species) is a finalist-level government-backed project developed in collaboration with INCOIS – Indian National Centre for Ocean Information Services, Hyderabad. It uses deep learning to detect and classify marine species from netted fish data, empowering sustainable fishing and aiding

Classifying the paintings of the 50 greatest artists in history using a convolutional neural network

LatentAugment, is a new data augmentation method for training any deep models. It navigates GAN's latent space to increase the diversity and quality of generated samples and enhance their effectiveness for DA purposes.

A deep learning model for classifying Indian medicinal plant leaves using ResNet50 and Keras. Trained on 80+ plant species with image augmentation and transfer learning.

Boosting Document Intelligence

SimplePY is a collection of Python scripts and utilities for file manipulation tailored for AI and ML projects. This repository aims to simplify the handling of datasets, model outputs, and other files commonly encountered in AI and machine learning workflows.

This repository contains Python script to augment image datas.

PlateDetectAI is a project using a custom YOLOv3 model for accurate vehicle number plate detection, featuring advanced image augmentation and custom loss functions.