Swin-Transformer-Semantic-Segmentation
mmsegmentation
| Swin-Transformer-Semantic-Segmentation | mmsegmentation | |
|---|---|---|
| 1 | 7 | |
| 1,081 | 7,414 | |
| 0.0% | 1.8% | |
| 0.0 | 8.2 | |
| over 1 year ago | 6 days ago | |
| Python | Python | |
| Apache License 2.0 | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
Swin-Transformer-Semantic-Segmentation
-
[P] Code and pretrained models for Swin Transformer are released (SOTA models on COCO and ADE20K)
Semantic segmentation on ADE20K: https://github.com/SwinTransformer/Swin-Transformer-Semantic-Segmentation
mmsegmentation
- [D] The MMSegmentation library from OpenMMLab appears to return the wrong results when computing basic image segmentation metrics such as the Jaccard index (IoU - intersection-over-union). It appears to compute recall (sensitivity) instead of IoU, which artificially inflates the performance metrics.
-
Is there any ML model out there for room surfaces detection? (ceiling, floor, windows)
Segmentation models trained on datasets like ADE20k could probably be used for that, because it has separate classes for these things iirc. https://github.com/open-mmlab/mmsegmentation should have suitable pretrained models available.
-
MMDeploy: Deploy All the Algorithms of OpenMMLab
MMSegmentation: OpenMMLab semantic segmentation toolbox and benchmark.
- Mmsegmentation - Openmmlab semantic segmentation toolbox and benchmark.
- Mmsegmentation – Openmmlab semantic segmentation toolbox and benchmark
-
Semantic Segmentation models
This repo is amazing: https://github.com/open-mmlab/mmsegmentation
-
What's A Simple Custom Segmentation Pipeline?
Mmsegmentation would be a good place to start for basic segmentation. They have lots of recent methods and pretained models you could fine-tune from. They also support quite a few datasets including VOC. There is a custom dataset format which looks straightforward to create.
What are some alternatives?
Swin-Transformer-Object-Detection - This is an official implementation for "Swin Transformer: Hierarchical Vision Transformer using Shifted Windows" on Object Detection and Instance Segmentation.
Pytorch-UNet - PyTorch implementation of the U-Net for image semantic segmentation with high quality images
labelme - Image Polygonal Annotation with Python (polygon, rectangle, circle, line, point and image-level flag annotation).
segmentation_models.pytorch - Segmentation models with pretrained backbones. PyTorch.
Swin-Transformer-Serve - Deploy Swin Transformer using TorchServe
Mask_RCNN - Mask R-CNN for object detection and instance segmentation on Keras and TensorFlow
Video-Swin-Transformer - This is an official implementation for "Video Swin Transformers".
face-parsing.PyTorch - Using modified BiSeNet for face parsing in PyTorch
PaddleClas - A treasure chest for visual classification and recognition powered by PaddlePaddle
PaddleSeg - Easy-to-use image segmentation library with awesome pre-trained model zoo, supporting wide-range of practical tasks in Semantic Segmentation, Interactive Segmentation, Panoptic Segmentation, Image Matting, 3D Segmentation, etc.
SeMask-Segmentation - [NIVT Workshop @ ICCV 2023] SeMask: Semantically Masked Transformers for Semantic Segmentation
labelme2coco - A lightweight package for converting your labelme annotations into COCO object detection format.