Name: Wenchao QI
Type: User
Company: Aerospace Information Research Institute, Chinese Academy of Sciences
Bio: Assistant Researcher at AIR, CAS, China. I'm mainly engaged in deep learning in Hyperspectral Computer Vision.
Location: No. 20, Datun Road, Chaoyang District, Beijing, China
Wenchao QI's Projects
2018 IEEE GRSS Data Fusion Contest
Classification of Hyperspectral images of Indian Pines dataset using a hybrid network of 3D and 2D CNN
Hyperspectral Image Classification using Deep Neural Network Architectures with Transfer Learning
A list of hyperspectral image super-solution resources collected by Junjun Jiang
Brief Introduction about Paper : Classification of Hyperspectral Imagery Using a New Fully Convolutional Neural Network
The official released demo of Tang Xu's group about the hyperspectral images classification
Hyperspectral Images Classification in Pytorch with multiple GPUs - Pavia University
3D Convolutional Adversial Autoencoder for Hyperspectral Classification
Code of paper "Deep Learning Classifiers for Hyperspectral Imaging: A Review"
Danfeng Hong, Lianru Gao, Renlong Hang, Bing Zhang, Jocelyn Chanussot. Deep Encoder-Decoder Networks for Classification of Hyperspectral and LiDAR Data, IEEE GRSL, 2020.
Danfeng Hong, Lianru Gao, Jing Yao, Bing Zhang, Antonio Plaza, Jocelyn Chanussot. Graph Convolutional Networks for Hyperspectral Image Classification, IEEE TGRS, 2020.
Semantic Segmentation Using Adversarial Networks - Project A
This project was done as a part of the Applied Machine Learning Course (COMP 551) at McGill University and was done in a group of 3 students. The goal of the project was that given an image which contains 2 single digit number, predict the sum of those single digits. The data-set consisted of 100,000 gray scale images which contained 2 single digits, these images were formed by combining two different images from the very famous MNIST Dataset. Four algorithms were applied on this dataset- Logistic Regression (LR), Support Vector Machine (SVM) Fully Connected Neural Network (NN) and Convolution Neural Network (CNN). We evaluated and discussed the results of these 4 algorithms on the dataset. It was observed that CNN performed best among the 4 algorithms which was not surprising as CNN is renowned to work well with images. My responsibility in this project included feature pre-processing and applying CNN. This project introduced me to the world of Deep Learning (read Keras, Tensorflows, GPU)
Implemented code for semi-automatic binary segmentation based on SLIC superpixels and graph-cuts.
With VAE and RNN (GRU or LSTM) I try to generate image sequences
pytorch Implementation of U-Net, R2U-Net, Attention U-Net, Attention R2U-Net.
A python library built to empower developers to build applications and systems with self-contained Computer Vision capabilities
该资源为作者在CSDN的撰写Python图像处理文章的支撑,主要是Python实现图像处理、图像识别、图像分类等算法代码实现,希望该资源对您有所帮助,一起加油。
Video Object Segmentation using Graph Neural Networks
Random Forest Regression of Chlorophyll-a concentration in Indian River Lagoon using Hyperspectral Satellite Imagery and in-situ measurements.
In-Place Activated BatchNorm for Memory-Optimized Training of DNNs
Unsupervised Intra-domain Adaptation for Semantic Segmentation through Self-Supervision (CVPR 2020)
Deep Prototypical Networks With Hybrid Residual Attention for Hyperspectral Image Classification, JSTARS, 2020
Kaggle competition solutions
DSTL Satellite Imagery Feature Detection Competition (68 out of 419)
Keras implementation of 3D Generative Adversarial Network.
It‘s my graduation design, which is designed to restore missing values in remote-sensing images.
Keras implementation of Deeplab v3+ with pretrained weights