Home

Awesome

PyTorch_UCMerced_LandUse

Dataset

UC Merced Land Use Dataset

Result

lr = 0.001, batch_size = 8

Epoch4100200
ResNet1845.48%86.19%91.90%
ResNet3444.05%79.52%79.52%

File structure

├── Images
│   ├── agricultural
│   ├── airplane
│   ├── baseballdiamond
│   ├── beach
│   ├── buildings
│   ├── chaparral
│   ├── denseresidential
│   ├── forest
│   ├── freeway
│   ├── golfcourse
│   ├── harbor
│   ├── intersection
│   ├── mediumresidential
│   ├── mobilehomepark
│   ├── overpass
│   ├── parkinglot
│   ├── river
│   ├── runway
│   ├── sparseresidential
│   ├── storagetanks
│   └── tenniscourt
├── README.md
├── checkpoint
│   ├── LeNet
│   │   └── ckpt.pth
│   └── ResNet
│       └── ckpt.pth
├── main.py
├── models
│   ├── __init__.py
│   ├── lenet.py
│   └── resnet.py
└── utils.py

Reference

  1. Yi Yang and Shawn Newsam, "Bag-Of-Visual-Words and Spatial Extensions for Land-Use Classification," ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (ACM GIS), 2010.