https://github.com/Ji-Xinyou/DIP-proj-DepthEstimation

Res-Unet for Depth Estimation
https://github.com/Ji-Xinyou/DIP-proj-DepthEstimation

Last synced: 3 months ago
JSON representation

Res-Unet for Depth Estimation

• Host: GitHub
• URL: https://github.com/Ji-Xinyou/DIP-proj-DepthEstimation
• Owner: xyjixyjixyji
• License: mit
• Created: 2021-11-23T08:54:12.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2021-12-13T03:10:31.000Z (almost 3 years ago)
• Last Synced: 2024-08-06T23:25:51.459Z (3 months ago)
• Language: Python
• Homepage:
• Size: 152 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-cs - @Ji-Xinyou, 2021 Fall

README

        
# DIP-proj-DepthEstimation

## Directories

* ./model

  * block.py

    * define the encoder(now only resnet50), mff, decoder and refinement block, which are modules of the model

  * model.py

    * define the model, utilizing the blocks in block.py

  * resnet_module.py

    * self-defined resnet for encoding blocks

      * credit to https://github.com/JunjH/

* nyu2_train

  * only part of the dataset, used to check whether the code is runnable

* load_data.py

  * load the data by making pairs of **path**, the output is a list of pair \[x\_tr\_path, y\_tr\_path\]

* loss.py

  * define the gradient computation and loss computation

* utils.py

  * currently, only loading and saving the params of model is defined

* train.py

  * define the training procedure