Classify the Scene of an Image and also provide Objects in the scene in audio file

Repository for training and evaluating various CNN based Classification Models

This GitHub repository focuses on an integrated approach to scene classification and image caption generation, aiming to improve the accuracy of scene evaluation in computer vision applications.

This research mini-project trains an expandable image classification system for place categorization which solves the closet-set limitation of convnets. The state-of-the-art Places365 convnet is trained using Places365 dataset with one vs all random forest classifier that outputs place labels.

A Deep Residual Autoencoder approach to colorize images.

"Deep Context-Aware Descreening and Rescreening of Halftone Images" paper implementation.

Some functions to preprocess Places365-Standard data set for deep learning tasks. (feed as data to CoarseNet, ObjectNet, EdgeNet, and DetailsNet)

VGG16 trained on Places365 model in Keras

The Places205 and Places365 GoogleNet Caffe models converted to PyTorch

A simple Convolutional Neural Network to find edges of images using Canny edge detector result as ground truth

Keras implementation of "Image Inpainting via Generative Multi-column Convolutional Neural Networks" paper published at NIPS 2018

Keras code and weights files for the VGG16-places365 and VGG16-hybrid1365 CNNs for scene classification