Pneumo.ai is a Streamlit web app for detecting pneumonia from X-rays, leveraging CNN variants like AlexNet, VGG, and ResNet18, with ResNet18 leading with an accuracy of 0.89. It employs early stopping callbacks for efficient model training, visualizes training using TensorBoard, and optimizes inference time with pruning and quantization techniques.

CNN-based classifiers for classifying audio samples into 13 distinct categories, such as laughter, car horn, dog bark, etc.

EthnicityPredictor-UTK is a PyTorch-based project that focuses on predicting ethnicities from facial images using state-of-the-art ResNet architectures. Leveraging the UTKfaces dataset, the model is trained to recognize diverse facial features and make accurate ethnicity predictions.

A pre-trained image classifier to classify dog breeds. Udacity Nanodegree project-1.

Efficient and Lightweight Ear Segmentation AI Model

Convolutional Neural Network for Face Recognition

HAM10000 Skin Lesion Classification

Automated Wildlife Monitoring Using Deep Learning

Using the GTZAN dataset to create a classifier that can classify what the music genre is playing!

Financial Time Series Forecasting using Deep Learning Techniques and Innovative Image Encoding Approaches

A Flask app for handwritten digit prediction.

Resnet-18 from scratch, using Pytorch

Light-weight Single Person Pose Estimator

Leaf Disease Detection: A project enabling farmers to identify plant diseases by scanning leaves, providing brief descriptions and suggesting supplements for treatment.

Deep Learning

Проект ML/DL определения сердечной недостаточности по векторной электрокардиографии

PyTorch implementation of Conditional Generative Adversarial Networks (cGAN) for image colorization of the MS COCO dataset

Text analysis using text-to-image conversion and deep neural models trained for image processing

Use yolov5 and ResNet18 for player tracking