Datasets, Transforms and Models specific to Computer Vision

In this paper I introduce a new oral lesion database that I created by scraping Images from research papers and dentists online. Next, I showcase a novel method for generating NxN pixel samples from human lips by isolating the labial fissure via image segmentation induced via the Kmean algorithm. I experiment with a variety of convolutional neur…

Convolutional Neural Network

Detect Fire Smoke with vgg and cnn

Image Classification for a City Dog Show Project Goal Improving your programming skills using Python. In this project, I use a created image classifier to identify dog breeds. It is for one of the submission for the udacity nanodegree program

Transfer learning on VGG16 using Keras with Caltech256 and Urban Tribes dataset. Dark knowledge in transfer learning.

使用keras实现show and tell。类似于https://github.com/oarriaga/neural_image_captioning ，但增加了Beam功能，可一次性给出beam_size个结果

🎨 Automatic Image Colorization using TensorFlow based on Residual Encoder Network

CNN distinguishing between men and women by photo

Binary classification of COVID-19 and non-COVID-19 chest X-rays

Funciones utilles para PyTorch.

Re-use the feature extractor of a model trained for object detection in a new model designed for Style Transfer

Pytorch implementations of each of the models described in the paper "Very Deep Convolutional Networks for Large-Scale Image Recognition" by Karen Simonyan, Andrew Zisserman

Machine Learning Practical - Coursework 2 Report: Analysing problems with the VGG deep neural network architectures (with 8 and 38 hidden layers) on the CIFAR100 dataset by monitoring gradient flow during training. And exploring solutions using batch normalization and residual connections.

Coding Neural Network architectures using PyTorch

A comprehensive collection of PyTorch implementations for the VGG (Visual Geometry Group) models

VGG model based on Tensorflow

Computer Vision models (include classification models and detection models) implemented by PyTorch

Image segmentation project. Two architectures implemented: VGG-16 + FCN-8 module and U-Net. For FCN-8, pre-trained weights are used from SSD300. Although it is designed for object detection, its feature extractor can be reused in another task involving similar classes. Linked article explains the full project.

A Two-Stage Deep Learning Approach for Mitosis Recognition