A deep learning-based classification system achieving 94% accuracy in distinguishing between diabetic retinopathy, glaucoma, and cataracts from normal eye images using fundus photography.

Trained a multiclass classifer network using cifar100 dataset

This repository contains Jupyter Notebook implementations of CNNs for image classification tasks that is Binary Image Classification, Multiclass Image classification. These notebooks were created during my learning journey with CNNs, covering key concepts such as data preprocessing, model building, training, evaluation, and performance metrics.

This project focuses on classifying distinct weather images using Convolutional Neural Networks (CNN) built from scratch and fine-tuning various state-of-the-art (SOTA) pre-trained models like AlexNet, ResNet50, VGG16, and MobileNet v3 Large. The models are trained and evaluated on a custom weather dataset, leveraging PyTorch for deep learning.

This project focuses on accurately classifying images of cats, dogs, and snakes using Convolutional Neural Networks (CNNs) in PyTorch. A custom CNN model was initially designed and trained, achieving strong classification performance. Additionally, state-of-the-art (SOTA) pre-trained image classification models such as AlexNet, ResNet50, and VGG16.

The Indian Bovine Classification project is an innovative web application designed to revolutionize the identification and study of Indian cattle breeds. Combining state-of-the-art image recognition technology with an intelligent chatbot system, this tool offers a comprehensive solution for classifying and learning about various Indian bovine breed

Building a CNN to identify hand written digits

Multiclass image classification using Convolutional Neural Network

This repository contains a deep learning model for skin cancer classification using the InceptionV3 architecture. The model was trained on the HAM10000 dataset and is designed with computational efficiency in mind. It was developed to be able to run on a CPU.

Implementation of V architecture with Vission Transformer for Image Segemntion Task

Basic classification algorithm for the MNIST dataset that uses computer vision fundamentals including simple NNs

This repository demonstrates a TensorFlow/Keras model for classifying MNIST handwritten digits. It covers data loading, normalization, model creation, training with early stopping, and evaluation. Includes visualizations of loss and accuracy over epochs and explains image shape considerations for predictions.

The project focuses on Identification of various Gemstone. The dataset consists of 87 classes.It shows the whole progress and model used to achieve final accuracy. You will gain knowledge of Computer Vision, The model used are CNN(Convolutional Neural Network), MobileNetV2 and VGGNet,The final model used was transfer learning with model MobileNetV2

The Bird Species Classifier is an application built using a Convolutional Neural Network (CNN) to classify images of birds into one of 525 different species. It allows users to upload an image of a bird and receive a prediction of the bird species. Along with analysing the performance of various optimising algorithms.

This is a project focused on identifying the presence of pneumonia in chest X-ray images. Each image can be classified into one of three categories: Bacterial Pneumonia, Viral Pneumonia, or Normal.

Computer vision deep learning project classifying 10 classes of food images with 90% accuracy, built with TensorFlow.

Computer vision deep learning project classifying 101 classes of food images with 80% accuracy, built with TensorFlow. Beats the baseline accuracy of 50.76% (Food101 paper) and 77.4% (DeepFood paper).

Skin Lesion Classifier using the ISIC 2018 Task 3 Dataset.