caltech101

Deep ConvNet Image Classifier based on Residual Network architecture trained on Caltech 101 Object Dataset

Replication of DeCAF paper's experiments for transfer learning

This repository contains the implementation of a fault detection system that detects and eliminates faulty products based on shape and color using a Convolutional Neural Network (CNN).

My Deep Learning Work

CNN-based image classification

Image Classification using Machine Learning, Neural Nets and CNNs.

Imperial College London EE4-62 Machine Learning for Computer Vision Coursework 1

Image recognition on CIFAR 10, CIFAR 100, Caltech 101 and Caltech 256 datasets. With the implementation of WideResNet, InceptionV3 and DenseNet neural networks.

Reimplementation of "VAE with a VampPrior" by Jakub M. Tomczak et al., as part of the DD2434 Machine Learning, Advanced Course at KTH

PyTorch Tutorials for several cases

Caltech-101 image classification using EfficientNet-B4 with 98.56% training accuracy and 93% validation accuracy.

Image Classification performed in HistogramData

CNN Image Identification

Pytorch Implementations of Neural Networks

This repo explores the effect of pre-trainig and experiments fine-tuning by applying different conbinations of hyperparameters.

A clean, reproducible, and well-structured classical machine learning pipeline for multi-class image classification on the Caltech101 dataset.

Content-Based Image Retrieval Using CNN and Hash