# Cell Detection
[](https://pepy.tech/project/celldetection)
[](https://github.com/FZJ-INM1-BDA/celldetection/actions?query=workflow%3ATest)
[](https://pypi.org/project/celldetection/)
[](https://celldetection.readthedocs.io/en/latest/?badge=latest)
[](https://zenodo.org/badge/latestdoi/349111085)
## ⭐ Showcase
###### NeurIPS 22 Cell Segmentation Competition
*https://openreview.net/forum?id=YtgRjBw-7GJ*
###### Nuclei of U2OS cells in a chemical screen
*https://bbbc.broadinstitute.org/BBBC039 (CC0)*
###### P. vivax (malaria) infected human blood
*https://bbbc.broadinstitute.org/BBBC041 (CC BY-NC-SA 3.0)*
## 🛠 Install
Make sure you have [PyTorch](https://pytorch.org/get-started/locally/) installed.
### PyPI
```
pip install -U celldetection
```
### GitHub
```
pip install git+https://github.com/FZJ-INM1-BDA/celldetection.git
```
## 💾 Trained models
```python
model = cd.fetch_model(model_name, check_hash=True)
```
| model name | training data | link |
|---------------------------------------------|----------------------------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------------:|
| `ginoro_CpnResNeXt101UNet-fbe875f1a3e5ce2c` | BBBC039, BBBC038, Omnipose, Cellpose, Sartorius - Cell Instance Segmentation, Livecell, NeurIPS 22 CellSeg Challenge | [🔗](https://celldetection.org/torch/models/ginoro_CpnResNeXt101UNet-fbe875f1a3e5ce2c.pt) |
<details>
<summary style="font-weight: bold; color: #888888">Run a demo with a pretrained model</summary>
```python
import torch, cv2, celldetection as cd
from skimage.data import coins
from matplotlib import pyplot as plt
# Load pretrained model
device = 'cuda' if torch.cuda.is_available() else 'cpu'
model = cd.fetch_model('ginoro_CpnResNeXt101UNet-fbe875f1a3e5ce2c', check_hash=True).to(device)
model.eval()
# Load input
img = coins()
img = cv2.cvtColor(img, cv2.COLOR_GRAY2RGB)
print(img.dtype, img.shape, (img.min(), img.max()))
# Run model
with torch.no_grad():
x = cd.to_tensor(img, transpose=True, device=device, dtype=torch.float32)
x = x / 255 # ensure 0..1 range
x = x[None] # add batch dimension: Tensor[3, h, w] -> Tensor[1, 3, h, w]
y = model(x)
# Show results for each batch item
contours = y['contours']
for n in range(len(x)):
cd.imshow_row(x[n], x[n], figsize=(16, 9), titles=('input', 'contours'))
cd.plot_contours(contours[n])
plt.show()
```
</details>
## 🔬 Architectures
```python
import celldetection as cd
```
<details>
<summary style="font-weight: bold; color: #888888">Contour Proposal Networks</summary>
- [`cd.models.CPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CPN)
- [`cd.models.CpnU22`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnU22)
- [`cd.models.CPNCore`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CPNCore)
- [`cd.models.CpnResUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResUNet)
- [`cd.models.CpnSlimU22`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnSlimU22)
- [`cd.models.CpnWideU22`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnWideU22)
- [`cd.models.CpnResNet18FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet18FPN)
- [`cd.models.CpnResNet34FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet34FPN)
- [`cd.models.CpnResNet50FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet50FPN)
- [`cd.models.CpnResNeXt50FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNeXt50FPN)
- [`cd.models.CpnResNet101FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet101FPN)
- [`cd.models.CpnResNet152FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet152FPN)
- [`cd.models.CpnResNet18UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet18UNet)
- [`cd.models.CpnResNet34UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet34UNet)
- [`cd.models.CpnResNet50UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet50UNet)
- [`cd.models.CpnResNeXt101FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNeXt101FPN)
- [`cd.models.CpnResNeXt152FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNeXt152FPN)
- [`cd.models.CpnResNeXt50UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNeXt50UNet)
- [`cd.models.CpnResNet101UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet101UNet)
- [`cd.models.CpnResNet152UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet152UNet)
- [`cd.models.CpnResNeXt101UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNeXt101UNet)
- [`cd.models.CpnResNeXt152UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNeXt152UNet)
- [`cd.models.CpnWideResNet50FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnWideResNet50FPN)
- [`cd.models.CpnWideResNet101FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnWideResNet101FPN)
- [`cd.models.CpnMobileNetV3LargeFPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnMobileNetV3LargeFPN)
- [`cd.models.CpnMobileNetV3SmallFPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnMobileNetV3SmallFPN)
</details>
<details>
<summary style="font-weight: bold; color: #888888">PyTorch Image Models (timm)</summary>
Also have a look at [Timm Documentation](https://huggingface.co/docs/timm/index).
```python
import timm
timm.list_models(filter='*') # explore available models
```
- [`cd.models.CpnTimmMaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnTimmMaNet)
- [`cd.models.CpnTimmUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnTimmUNet)
- [`cd.models.TimmEncoder`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.timmodels.TimmEncoder)
- [`cd.models.TimmFPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.TimmFPN)
- [`cd.models.TimmMaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.manet.TimmMaNet)
- [`cd.models.TimmUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.TimmUNet)
</details>
<details>
<summary style="font-weight: bold; color: #888888">Segmentation Models PyTorch (smp)</summary>
```python
import segmentation_models_pytorch as smp
smp.encoders.get_encoder_names() # explore available models
```
```python
encoder = cd.models.SmpEncoder(encoder_name='mit_b5', pretrained='imagenet')
```
Find a list of [Smp Encoders](https://smp.readthedocs.io/en/latest/encoders.html) in the `smp` documentation.
- [`cd.models.CpnSmpMaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnSmpMaNet)
- [`cd.models.CpnSmpUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnSmpUNet)
- [`cd.models.SmpEncoder`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.smp.SmpEncoder)
- [`cd.models.SmpFPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.SmpFPN)
- [`cd.models.SmpMaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.manet.SmpMaNet)
- [`cd.models.SmpUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.SmpUNet)
</details>
<details>
<summary style="font-weight: bold; color: #888888">U-Nets</summary>
```python
# U-Nets are available in 2D and 3D
import celldetection as cd
model = cd.models.ResNeXt50UNet(in_channels=3, out_channels=1, nd=3)
```
- [`cd.models.U22`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.U22)
- [`cd.models.U17`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.U17)
- [`cd.models.U12`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.U12)
- [`cd.models.UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.UNet)
- [`cd.models.WideU22`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.WideU22)
- [`cd.models.SlimU22`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.SlimU22)
- [`cd.models.ResUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResUNet)
- [`cd.models.UNetEncoder`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.UNetEncoder)
- [`cd.models.ResNet50UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNet50UNet)
- [`cd.models.ResNet18UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNet18UNet)
- [`cd.models.ResNet34UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNet34UNet)
- [`cd.models.ResNet152UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNet152UNet)
- [`cd.models.ResNet101UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNet101UNet)
- [`cd.models.ResNeXt50UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNeXt50UNet)
- [`cd.models.ResNeXt152UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNeXt152UNet)
- [`cd.models.ResNeXt101UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNeXt101UNet)
- [`cd.models.WideResNet50UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.WideResNet50UNet)
- [`cd.models.WideResNet101UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.WideResNet101UNet)
- [`cd.models.MobileNetV3SmallUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.MobileNetV3SmallUNet)
- [`cd.models.MobileNetV3LargeUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.MobileNetV3LargeUNet)
</details>
<details>
<summary style="font-weight: bold; color: #888888">MA-Nets</summary>
```python
# Many MA-Nets are available in 2D and 3D
import celldetection as cd
encoder = cd.models.ConvNeXtSmall(in_channels=3, nd=3)
model = cd.models.MaNet(encoder, out_channels=1, nd=3)
```
- [`cd.models.MaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.manet.MaNet)
- [`cd.models.SmpMaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.manet.SmpMaNet)
- [`cd.models.TimmMaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.manet.TimmMaNet)
</details>
<details>
<summary style="font-weight: bold; color: #888888">Feature Pyramid Networks</summary>
- [`cd.models.FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.FPN)
- [`cd.models.ResNet18FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNet18FPN)
- [`cd.models.ResNet34FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNet34FPN)
- [`cd.models.ResNet50FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNet50FPN)
- [`cd.models.ResNeXt50FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNeXt50FPN)
- [`cd.models.ResNet101FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNet101FPN)
- [`cd.models.ResNet152FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNet152FPN)
- [`cd.models.ResNeXt101FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNeXt101FPN)
- [`cd.models.ResNeXt152FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNeXt152FPN)
- [`cd.models.WideResNet50FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.WideResNet50FPN)
- [`cd.models.WideResNet101FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.WideResNet101FPN)
- [`cd.models.MobileNetV3LargeFPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.MobileNetV3LargeFPN)
- [`cd.models.MobileNetV3SmallFPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.MobileNetV3SmallFPN)
</details>
<details>
<summary style="font-weight: bold; color: #888888">ConvNeXt Networks</summary>
```python
# ConvNeXt Networks are available in 2D and 3D
import celldetection as cd
model = cd.models.ConvNeXtSmall(in_channels=3, nd=3)
```
- [`cd.models.ConvNeXt`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.convnext.MaNet)
- [`cd.models.ConvNeXtTiny`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.convnext.ConvNeXtTiny)
- [`cd.models.ConvNeXtSmall`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.convnext.ConvNeXtSmall)
- [`cd.models.ConvNeXtBase`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.convnext.ConvNeXtBase)
- [`cd.models.ConvNeXtLarge`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.convnext.ConvNeXtLarge)
</details>
<details>
<summary style="font-weight: bold; color: #888888">Residual Networks</summary>
```python
# Residual Networks are available in 2D and 3D
import celldetection as cd
model = cd.models.ResNet50(in_channels=3, nd=3)
```
- [`cd.models.ResNet18`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNet18)
- [`cd.models.ResNet34`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNet34)
- [`cd.models.ResNet50`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNet50)
- [`cd.models.ResNet101`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNet101)
- [`cd.models.ResNet152`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNet152)
- [`cd.models.WideResNet50_2`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.WideResNet50_2)
- [`cd.models.ResNeXt50_32x4d`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNeXt50_32x4d)
- [`cd.models.WideResNet101_2`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.WideResNet101_2)
- [`cd.models.ResNeXt101_32x8d`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNeXt101_32x8d)
- [`cd.models.ResNeXt152_32x8d`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNeXt152_32x8d)
</details>
<details>
<summary style="font-weight: bold; color: #888888">Mobile Networks</summary>
- [`cd.models.MobileNetV3Large`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.mobilenetv3.MobileNetV3Large)
- [`cd.models.MobileNetV3Small`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.mobilenetv3.MobileNetV3Small)
</details>
## 🐳 Docker
Find us on Docker Hub: https://hub.docker.com/r/ericup/celldetection
You can pull the latest version of `celldetection` via:
```
docker pull ericup/celldetection:latest
```
<details>
<summary style="font-weight: bold; color: #888888">CPN inference via Docker with GPU</summary>
```
docker run --rm \
-v $PWD/docker/outputs:/outputs/ \
-v $PWD/docker/inputs/:/inputs/ \
-v $PWD/docker/models/:/models/ \
--gpus="device=0" \
celldetection:latest /bin/bash -c \
"python cpn_inference.py --tile_size=1024 --stride=768 --precision=32-true"
```
</details>
<details>
<summary style="font-weight: bold; color: #888888">CPN inference via Docker with CPU</summary>
```
docker run --rm \
-v $PWD/docker/outputs:/outputs/ \
-v $PWD/docker/inputs/:/inputs/ \
-v $PWD/docker/models/:/models/ \
celldetection:latest /bin/bash -c \
"python cpn_inference.py --tile_size=1024 --stride=768 --precision=32-true --accelerator=cpu"
```
</details>
### Apptainer
You can also pull our Docker images for the use with [Apptainer](https://apptainer.org/) (formerly [Singularity](https://github.com/apptainer/singularity)) with this command:
```
apptainer pull --dir . --disable-cache docker://ericup/celldetection:latest
```
## 🤗 Hugging Face Spaces
Find us on Hugging Face and upload your own images for segmentation: https://huggingface.co/spaces/ericup/celldetection
There's also an API (Python & JavaScript), allowing you to utilize community GPUs (currently Nvidia A100) remotely!
<details>
<summary style="font-weight: bold; color: #888888">Hugging Face API</summary>
### Python
```python
from gradio_client import Client
# Define inputs (local filename or URL)
inputs = 'https://raw.githubusercontent.com/scikit-image/scikit-image/main/skimage/data/coins.png'
# Set up client
client = Client("ericup/celldetection")
# Predict
overlay_filename, img_filename, h5_filename, csv_filename = client.predict(
inputs, # str: Local filepath or URL of your input image
# Model name
'ginoro_CpnResNeXt101UNet-fbe875f1a3e5ce2c',
# Custom Score Threshold (numeric value between 0 and 1)
False, .9, # bool: Whether to use custom setting; float: Custom setting
# Custom NMS Threshold
False, .3142, # bool: Whether to use custom setting; float: Custom setting
# Custom Number of Sample Points
False, 128, # bool: Whether to use custom setting; int: Custom setting
# Overlapping objects
True, # bool: Whether to allow overlapping objects
# API name (keep as is)
api_name="/predict"
)
# Example usage: Code below only shows how to use the results
from matplotlib import pyplot as plt
import celldetection as cd
import pandas as pd
# Read results from local temporary files
img = imread(img_filename)
overlay = imread(overlay_filename) # random colors per instance; transparent overlap
properties = pd.read_csv(csv_filename)
contours, scores, label_image = cd.from_h5(h5_filename, 'contours', 'scores', 'labels')
# Optionally display overlay
cd.imshow_row(img, img, figsize=(16, 9))
cd.imshow(overlay)
plt.show()
# Optionally display contours with text
cd.imshow_row(img, img, figsize=(16, 9))
cd.plot_contours(contours, texts=['score: %d%%\narea: %d' % s for s in zip((scores * 100).round(), properties.area)])
plt.show()
```
### Javascript
```javascript
import { client } from "@gradio/client";
const response_0 = await fetch("https://raw.githubusercontent.com/scikit-image/scikit-image/main/skimage/data/coins.png");
const exampleImage = await response_0.blob();
const app = await client("ericup/celldetection");
const result = await app.predict("/predict", [
exampleImage, // blob: Your input image
// Model name (hosted model or URL)
"ginoro_CpnResNeXt101UNet-fbe875f1a3e5ce2c",
// Custom Score Threshold (numeric value between 0 and 1)
false, .9, // bool: Whether to use custom setting; float: Custom setting
// Custom NMS Threshold
false, .3142, // bool: Whether to use custom setting; float: Custom setting
// Custom Number of Sample Points
false, 128, // bool: Whether to use custom setting; int: Custom setting
// Overlapping objects
true, // bool: Whether to allow overlapping objects
// API name (keep as is)
api_name="/predict"
]);
```
</details>
## 🧑💻 Napari Plugin
Find our Napari Plugin here: https://github.com/FZJ-INM1-BDA/celldetection-napari </br>
Find out more about Napari here: https://napari.org
You can install it via pip:
```
pip install git+https://github.com/FZJ-INM1-BDA/celldetection-napari.git
```
## 🏆 Awards
- [NeurIPS 2022 Cell Segmentation Challenge](https://neurips22-cellseg.grand-challenge.org/): Winner Finalist Award
## 📝 Citing
If you find this work useful, please consider giving a **star** ⭐️ and **citation**:
```
@article{UPSCHULTE2022102371,
title = {Contour proposal networks for biomedical instance segmentation},
journal = {Medical Image Analysis},
volume = {77},
pages = {102371},
year = {2022},
issn = {1361-8415},
doi = {https://doi.org/10.1016/j.media.2022.102371},
url = {https://www.sciencedirect.com/science/article/pii/S136184152200024X},
author = {Eric Upschulte and Stefan Harmeling and Katrin Amunts and Timo Dickscheid},
keywords = {Cell detection, Cell segmentation, Object detection, CPN},
}
```
## 🔗 Links
- [Article (sciencedirect)](https://www.sciencedirect.com/science/article/pii/S136184152200024X "Contour Proposal Networks for Biomedical Instance Segmentation")
- [PDF (sciencedirect)](https://www.sciencedirect.com/sdfe/reader/pii/S136184152200024X/pdf "Contour Proposal Networks for Biomedical Instance Segmentation")
- [PyPI](https://pypi.org/project/celldetection/ "CellDetection")
- [Documentation](https://docs.celldetection.org "Documentation")
## 🧑🔬 Thanks!
[](https://github.com/FZJ-INM1-BDA/celldetection/stargazers)
[](https://github.com/FZJ-INM1-BDA/celldetection/network/members)
Raw data
{
"_id": null,
"home_page": "https://celldetection.org",
"name": "CellDetection",
"maintainer": null,
"docs_url": null,
"requires_python": null,
"maintainer_email": null,
"keywords": "cell, detection, object, segmentation, pytorch, cpn, contour, proposal, network, deep, learning, unet, fzj, julich, juelich, ai",
"author": "Eric Upschulte",
"author_email": "e.upschulte@fz-juelich.de",
"download_url": "https://files.pythonhosted.org/packages/3e/4d/110e19ea629c1e1702d0f3a65d977dabcb2e5d9c652658961c64fe99955d/celldetection-0.4.8.tar.gz",
"platform": null,
"description": "# Cell Detection\n\n[](https://pepy.tech/project/celldetection)\n[](https://github.com/FZJ-INM1-BDA/celldetection/actions?query=workflow%3ATest)\n[](https://pypi.org/project/celldetection/)\n[](https://celldetection.readthedocs.io/en/latest/?badge=latest)\n[](https://zenodo.org/badge/latestdoi/349111085)\n\n## \u2b50 Showcase\n\n###### NeurIPS 22 Cell Segmentation Competition\n\n![neurips22](https://raw.githubusercontent.com/FZJ-INM1-BDA/celldetection/main/assets/neurips-cellseg-demo.png \"NeurIPS 22 Cell Segmentation Competition - Find more information here: https://neurips.cc/Conferences/2022/CompetitionTrack\")\n*https://openreview.net/forum?id=YtgRjBw-7GJ*\n\n###### Nuclei of U2OS cells in a chemical screen\n\n![bbbc039](https://raw.githubusercontent.com/FZJ-INM1-BDA/celldetection/main/assets/bbbc039-cpn-u22-demo.png \"BBBC039 demo with CpnU22 - Find the dataset here: https://bbbc.broadinstitute.org/BBBC039\")\n*https://bbbc.broadinstitute.org/BBBC039 (CC0)*\n\n###### P. vivax (malaria) infected human blood\n\n![bbbc041](https://raw.githubusercontent.com/FZJ-INM1-BDA/celldetection/main/assets/bbbc041-cpn-u22-demo.png \"BBBC041 demo with CpnU22 - Find the dataset here: https://bbbc.broadinstitute.org/BBBC041\")\n*https://bbbc.broadinstitute.org/BBBC041 (CC BY-NC-SA 3.0)*\n\n## \ud83d\udee0 Install\n\nMake sure you have [PyTorch](https://pytorch.org/get-started/locally/) installed.\n\n### PyPI\n\n```\npip install -U celldetection\n```\n\n### GitHub\n\n```\npip install git+https://github.com/FZJ-INM1-BDA/celldetection.git\n```\n\n## \ud83d\udcbe Trained models\n\n```python\nmodel = cd.fetch_model(model_name, check_hash=True)\n```\n\n| model name | training data | link |\n|---------------------------------------------|----------------------------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------------:| \n| `ginoro_CpnResNeXt101UNet-fbe875f1a3e5ce2c` | BBBC039, BBBC038, Omnipose, Cellpose, Sartorius - Cell Instance Segmentation, Livecell, NeurIPS 22 CellSeg Challenge | [\ud83d\udd17](https://celldetection.org/torch/models/ginoro_CpnResNeXt101UNet-fbe875f1a3e5ce2c.pt) |\n\n<details>\n <summary style=\"font-weight: bold; color: #888888\">Run a demo with a pretrained model</summary>\n\n```python\nimport torch, cv2, celldetection as cd\nfrom skimage.data import coins\nfrom matplotlib import pyplot as plt\n\n# Load pretrained model\ndevice = 'cuda' if torch.cuda.is_available() else 'cpu'\nmodel = cd.fetch_model('ginoro_CpnResNeXt101UNet-fbe875f1a3e5ce2c', check_hash=True).to(device)\nmodel.eval()\n\n# Load input\nimg = coins()\nimg = cv2.cvtColor(img, cv2.COLOR_GRAY2RGB)\nprint(img.dtype, img.shape, (img.min(), img.max()))\n\n# Run model\nwith torch.no_grad():\n x = cd.to_tensor(img, transpose=True, device=device, dtype=torch.float32)\n x = x / 255 # ensure 0..1 range\n x = x[None] # add batch dimension: Tensor[3, h, w] -> Tensor[1, 3, h, w]\n y = model(x)\n\n# Show results for each batch item\ncontours = y['contours']\nfor n in range(len(x)):\n cd.imshow_row(x[n], x[n], figsize=(16, 9), titles=('input', 'contours'))\n cd.plot_contours(contours[n])\n plt.show()\n```\n\n</details>\n\n## \ud83d\udd2c Architectures\n\n```python\nimport celldetection as cd\n```\n\n<details>\n <summary style=\"font-weight: bold; color: #888888\">Contour Proposal Networks</summary>\n\n- [`cd.models.CPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CPN)\n- [`cd.models.CpnU22`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnU22)\n- [`cd.models.CPNCore`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CPNCore)\n- [`cd.models.CpnResUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResUNet)\n- [`cd.models.CpnSlimU22`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnSlimU22)\n- [`cd.models.CpnWideU22`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnWideU22)\n- [`cd.models.CpnResNet18FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet18FPN)\n- [`cd.models.CpnResNet34FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet34FPN)\n- [`cd.models.CpnResNet50FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet50FPN)\n- [`cd.models.CpnResNeXt50FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNeXt50FPN)\n- [`cd.models.CpnResNet101FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet101FPN)\n- [`cd.models.CpnResNet152FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet152FPN)\n- [`cd.models.CpnResNet18UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet18UNet)\n- [`cd.models.CpnResNet34UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet34UNet)\n- [`cd.models.CpnResNet50UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet50UNet)\n- [`cd.models.CpnResNeXt101FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNeXt101FPN)\n- [`cd.models.CpnResNeXt152FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNeXt152FPN)\n- [`cd.models.CpnResNeXt50UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNeXt50UNet)\n- [`cd.models.CpnResNet101UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet101UNet)\n- [`cd.models.CpnResNet152UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNet152UNet)\n- [`cd.models.CpnResNeXt101UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNeXt101UNet)\n- [`cd.models.CpnResNeXt152UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnResNeXt152UNet)\n- [`cd.models.CpnWideResNet50FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnWideResNet50FPN)\n- [`cd.models.CpnWideResNet101FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnWideResNet101FPN)\n- [`cd.models.CpnMobileNetV3LargeFPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnMobileNetV3LargeFPN)\n- [`cd.models.CpnMobileNetV3SmallFPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnMobileNetV3SmallFPN)\n\n</details>\n\n<details>\n <summary style=\"font-weight: bold; color: #888888\">PyTorch Image Models (timm)</summary>\n\nAlso have a look at [Timm Documentation](https://huggingface.co/docs/timm/index).\n\n```python\nimport timm\n\ntimm.list_models(filter='*') # explore available models\n```\n\n- [`cd.models.CpnTimmMaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnTimmMaNet)\n- [`cd.models.CpnTimmUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnTimmUNet)\n- [`cd.models.TimmEncoder`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.timmodels.TimmEncoder)\n- [`cd.models.TimmFPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.TimmFPN)\n- [`cd.models.TimmMaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.manet.TimmMaNet)\n- [`cd.models.TimmUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.TimmUNet)\n\n</details>\n\n<details>\n <summary style=\"font-weight: bold; color: #888888\">Segmentation Models PyTorch (smp)</summary>\n\n```python\nimport segmentation_models_pytorch as smp\n\nsmp.encoders.get_encoder_names() # explore available models\n```\n\n```python\nencoder = cd.models.SmpEncoder(encoder_name='mit_b5', pretrained='imagenet')\n```\n\nFind a list of [Smp Encoders](https://smp.readthedocs.io/en/latest/encoders.html) in the `smp` documentation.\n\n- [`cd.models.CpnSmpMaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnSmpMaNet)\n- [`cd.models.CpnSmpUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.cpn.CpnSmpUNet)\n- [`cd.models.SmpEncoder`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.smp.SmpEncoder)\n- [`cd.models.SmpFPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.SmpFPN)\n- [`cd.models.SmpMaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.manet.SmpMaNet)\n- [`cd.models.SmpUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.SmpUNet)\n\n</details>\n\n<details>\n <summary style=\"font-weight: bold; color: #888888\">U-Nets</summary>\n\n```python\n# U-Nets are available in 2D and 3D\nimport celldetection as cd\n\nmodel = cd.models.ResNeXt50UNet(in_channels=3, out_channels=1, nd=3)\n```\n\n- [`cd.models.U22`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.U22)\n- [`cd.models.U17`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.U17)\n- [`cd.models.U12`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.U12)\n- [`cd.models.UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.UNet)\n- [`cd.models.WideU22`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.WideU22)\n- [`cd.models.SlimU22`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.SlimU22)\n- [`cd.models.ResUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResUNet)\n- [`cd.models.UNetEncoder`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.UNetEncoder)\n- [`cd.models.ResNet50UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNet50UNet)\n- [`cd.models.ResNet18UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNet18UNet)\n- [`cd.models.ResNet34UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNet34UNet)\n- [`cd.models.ResNet152UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNet152UNet)\n- [`cd.models.ResNet101UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNet101UNet)\n- [`cd.models.ResNeXt50UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNeXt50UNet)\n- [`cd.models.ResNeXt152UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNeXt152UNet)\n- [`cd.models.ResNeXt101UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.ResNeXt101UNet)\n- [`cd.models.WideResNet50UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.WideResNet50UNet)\n- [`cd.models.WideResNet101UNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.WideResNet101UNet)\n- [`cd.models.MobileNetV3SmallUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.MobileNetV3SmallUNet)\n- [`cd.models.MobileNetV3LargeUNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.unet.MobileNetV3LargeUNet)\n\n</details>\n\n<details>\n <summary style=\"font-weight: bold; color: #888888\">MA-Nets</summary>\n\n```python\n# Many MA-Nets are available in 2D and 3D\nimport celldetection as cd\n\nencoder = cd.models.ConvNeXtSmall(in_channels=3, nd=3)\nmodel = cd.models.MaNet(encoder, out_channels=1, nd=3)\n```\n\n- [`cd.models.MaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.manet.MaNet)\n- [`cd.models.SmpMaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.manet.SmpMaNet)\n- [`cd.models.TimmMaNet`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.manet.TimmMaNet)\n\n</details>\n\n<details>\n <summary style=\"font-weight: bold; color: #888888\">Feature Pyramid Networks</summary>\n\n- [`cd.models.FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.FPN)\n- [`cd.models.ResNet18FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNet18FPN)\n- [`cd.models.ResNet34FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNet34FPN)\n- [`cd.models.ResNet50FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNet50FPN)\n- [`cd.models.ResNeXt50FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNeXt50FPN)\n- [`cd.models.ResNet101FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNet101FPN)\n- [`cd.models.ResNet152FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNet152FPN)\n- [`cd.models.ResNeXt101FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNeXt101FPN)\n- [`cd.models.ResNeXt152FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.ResNeXt152FPN)\n- [`cd.models.WideResNet50FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.WideResNet50FPN)\n- [`cd.models.WideResNet101FPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.WideResNet101FPN)\n- [`cd.models.MobileNetV3LargeFPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.MobileNetV3LargeFPN)\n- [`cd.models.MobileNetV3SmallFPN`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.fpn.MobileNetV3SmallFPN)\n\n</details>\n\n<details>\n <summary style=\"font-weight: bold; color: #888888\">ConvNeXt Networks</summary>\n\n```python\n# ConvNeXt Networks are available in 2D and 3D\nimport celldetection as cd\n\nmodel = cd.models.ConvNeXtSmall(in_channels=3, nd=3)\n```\n\n- [`cd.models.ConvNeXt`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.convnext.MaNet)\n- [`cd.models.ConvNeXtTiny`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.convnext.ConvNeXtTiny)\n- [`cd.models.ConvNeXtSmall`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.convnext.ConvNeXtSmall)\n- [`cd.models.ConvNeXtBase`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.convnext.ConvNeXtBase)\n- [`cd.models.ConvNeXtLarge`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.convnext.ConvNeXtLarge)\n\n</details>\n\n<details>\n <summary style=\"font-weight: bold; color: #888888\">Residual Networks</summary>\n\n```python\n# Residual Networks are available in 2D and 3D\nimport celldetection as cd\n\nmodel = cd.models.ResNet50(in_channels=3, nd=3)\n```\n\n- [`cd.models.ResNet18`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNet18)\n- [`cd.models.ResNet34`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNet34)\n- [`cd.models.ResNet50`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNet50)\n- [`cd.models.ResNet101`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNet101)\n- [`cd.models.ResNet152`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNet152)\n- [`cd.models.WideResNet50_2`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.WideResNet50_2)\n- [`cd.models.ResNeXt50_32x4d`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNeXt50_32x4d)\n- [`cd.models.WideResNet101_2`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.WideResNet101_2)\n- [`cd.models.ResNeXt101_32x8d`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNeXt101_32x8d)\n- [`cd.models.ResNeXt152_32x8d`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.resnet.ResNeXt152_32x8d)\n\n</details>\n\n<details>\n <summary style=\"font-weight: bold; color: #888888\">Mobile Networks</summary>\n\n- [`cd.models.MobileNetV3Large`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.mobilenetv3.MobileNetV3Large)\n- [`cd.models.MobileNetV3Small`](https://docs.celldetection.org/en/latest/celldetection.models.html#celldetection.models.mobilenetv3.MobileNetV3Small)\n\n</details>\n\n## \ud83d\udc33 Docker\n\nFind us on Docker Hub: https://hub.docker.com/r/ericup/celldetection\n\nYou can pull the latest version of `celldetection` via:\n```\ndocker pull ericup/celldetection:latest\n```\n\n<details>\n <summary style=\"font-weight: bold; color: #888888\">CPN inference via Docker with GPU</summary>\n\n```\ndocker run --rm \\\n -v $PWD/docker/outputs:/outputs/ \\\n -v $PWD/docker/inputs/:/inputs/ \\\n -v $PWD/docker/models/:/models/ \\\n --gpus=\"device=0\" \\\n celldetection:latest /bin/bash -c \\\n \"python cpn_inference.py --tile_size=1024 --stride=768 --precision=32-true\"\n```\n</details>\n<details>\n <summary style=\"font-weight: bold; color: #888888\">CPN inference via Docker with CPU</summary>\n\n```\ndocker run --rm \\\n -v $PWD/docker/outputs:/outputs/ \\\n -v $PWD/docker/inputs/:/inputs/ \\\n -v $PWD/docker/models/:/models/ \\\n celldetection:latest /bin/bash -c \\\n \"python cpn_inference.py --tile_size=1024 --stride=768 --precision=32-true --accelerator=cpu\"\n```\n</details>\n\n\n\n### Apptainer\n\nYou can also pull our Docker images for the use with [Apptainer](https://apptainer.org/) (formerly [Singularity](https://github.com/apptainer/singularity)) with this command:\n\n```\napptainer pull --dir . --disable-cache docker://ericup/celldetection:latest\n```\n\n\n## \ud83e\udd17 Hugging Face Spaces\n\nFind us on Hugging Face and upload your own images for segmentation: https://huggingface.co/spaces/ericup/celldetection\n\nThere's also an API (Python & JavaScript), allowing you to utilize community GPUs (currently Nvidia A100) remotely!\n\n<details>\n <summary style=\"font-weight: bold; color: #888888\">Hugging Face API</summary>\n\n### Python\n\n```python\nfrom gradio_client import Client\n\n# Define inputs (local filename or URL)\ninputs = 'https://raw.githubusercontent.com/scikit-image/scikit-image/main/skimage/data/coins.png'\n\n# Set up client\nclient = Client(\"ericup/celldetection\")\n\n# Predict\noverlay_filename, img_filename, h5_filename, csv_filename = client.predict(\n inputs, # str: Local filepath or URL of your input image\n \n # Model name\n 'ginoro_CpnResNeXt101UNet-fbe875f1a3e5ce2c',\n \n # Custom Score Threshold (numeric value between 0 and 1)\n False, .9, # bool: Whether to use custom setting; float: Custom setting\n \n # Custom NMS Threshold\n False, .3142, # bool: Whether to use custom setting; float: Custom setting\n \n # Custom Number of Sample Points\n False, 128, # bool: Whether to use custom setting; int: Custom setting\n \n # Overlapping objects\n True, # bool: Whether to allow overlapping objects\n \n # API name (keep as is)\n api_name=\"/predict\"\n)\n\n\n# Example usage: Code below only shows how to use the results\nfrom matplotlib import pyplot as plt\nimport celldetection as cd\nimport pandas as pd\n\n# Read results from local temporary files\nimg = imread(img_filename)\noverlay = imread(overlay_filename) # random colors per instance; transparent overlap\nproperties = pd.read_csv(csv_filename)\ncontours, scores, label_image = cd.from_h5(h5_filename, 'contours', 'scores', 'labels')\n\n# Optionally display overlay\ncd.imshow_row(img, img, figsize=(16, 9))\ncd.imshow(overlay)\nplt.show()\n\n# Optionally display contours with text\ncd.imshow_row(img, img, figsize=(16, 9))\ncd.plot_contours(contours, texts=['score: %d%%\\narea: %d' % s for s in zip((scores * 100).round(), properties.area)])\nplt.show()\n```\n\n### Javascript\n\n```javascript\nimport { client } from \"@gradio/client\";\n\nconst response_0 = await fetch(\"https://raw.githubusercontent.com/scikit-image/scikit-image/main/skimage/data/coins.png\");\nconst exampleImage = await response_0.blob();\n\t\t\t\t\t\t\nconst app = await client(\"ericup/celldetection\");\nconst result = await app.predict(\"/predict\", [\n exampleImage, // blob: Your input image\n \n // Model name (hosted model or URL)\n \"ginoro_CpnResNeXt101UNet-fbe875f1a3e5ce2c\",\n \n // Custom Score Threshold (numeric value between 0 and 1)\n false, .9, // bool: Whether to use custom setting; float: Custom setting\n \n // Custom NMS Threshold\n false, .3142, // bool: Whether to use custom setting; float: Custom setting\n \n // Custom Number of Sample Points\n false, 128, // bool: Whether to use custom setting; int: Custom setting\n \n // Overlapping objects\n true, // bool: Whether to allow overlapping objects\n \n // API name (keep as is)\n api_name=\"/predict\"\n]);\n```\n\n</details>\n\n## \ud83e\uddd1\u200d\ud83d\udcbb Napari Plugin\n\nFind our Napari Plugin here: https://github.com/FZJ-INM1-BDA/celldetection-napari </br>\nFind out more about Napari here: https://napari.org\n![bbbc039](https://raw.githubusercontent.com/FZJ-INM1-BDA/celldetection-napari/main/assets/coins-demo.png \"Napari Plugin\")\nYou can install it via pip:\n```\npip install git+https://github.com/FZJ-INM1-BDA/celldetection-napari.git\n```\n\n## \ud83c\udfc6 Awards\n\n- [NeurIPS 2022 Cell Segmentation Challenge](https://neurips22-cellseg.grand-challenge.org/): Winner Finalist Award\n\n## \ud83d\udcdd Citing\n\nIf you find this work useful, please consider giving a **star** \u2b50\ufe0f and **citation**:\n\n```\n@article{UPSCHULTE2022102371,\n title = {Contour proposal networks for biomedical instance segmentation},\n journal = {Medical Image Analysis},\n volume = {77},\n pages = {102371},\n year = {2022},\n issn = {1361-8415},\n doi = {https://doi.org/10.1016/j.media.2022.102371},\n url = {https://www.sciencedirect.com/science/article/pii/S136184152200024X},\n author = {Eric Upschulte and Stefan Harmeling and Katrin Amunts and Timo Dickscheid},\n keywords = {Cell detection, Cell segmentation, Object detection, CPN},\n}\n```\n\n## \ud83d\udd17 Links\n\n- [Article (sciencedirect)](https://www.sciencedirect.com/science/article/pii/S136184152200024X \"Contour Proposal Networks for Biomedical Instance Segmentation\")\n- [PDF (sciencedirect)](https://www.sciencedirect.com/sdfe/reader/pii/S136184152200024X/pdf \"Contour Proposal Networks for Biomedical Instance Segmentation\")\n- [PyPI](https://pypi.org/project/celldetection/ \"CellDetection\")\n- [Documentation](https://docs.celldetection.org \"Documentation\")\n\n## \ud83e\uddd1\u200d\ud83d\udd2c Thanks!\n\n[](https://github.com/FZJ-INM1-BDA/celldetection/stargazers)\n[](https://github.com/FZJ-INM1-BDA/celldetection/network/members)\n",
"bugtrack_url": null,
"license": "Apache License, Version 2.0",
"summary": "Cell Detection with PyTorch.",
"version": "0.4.8",
"project_urls": {
"Homepage": "https://celldetection.org"
},
"split_keywords": [
"cell",
" detection",
" object",
" segmentation",
" pytorch",
" cpn",
" contour",
" proposal",
" network",
" deep",
" learning",
" unet",
" fzj",
" julich",
" juelich",
" ai"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "bd5a0f232cbd586bff9b560a9d6043b0f7798f880ee403d66cadad8eed2811a2",
"md5": "b203d7b5d55c79ef02fc4df92ceaf651",
"sha256": "4426a78a95d1fe7f5fcda64fa7cce8dda4a89070249b2be26bc7e829e241bede"
},
"downloads": -1,
"filename": "CellDetection-0.4.8-py3-none-any.whl",
"has_sig": false,
"md5_digest": "b203d7b5d55c79ef02fc4df92ceaf651",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": null,
"size": 172498,
"upload_time": "2024-05-29T14:36:58",
"upload_time_iso_8601": "2024-05-29T14:36:58.452132Z",
"url": "https://files.pythonhosted.org/packages/bd/5a/0f232cbd586bff9b560a9d6043b0f7798f880ee403d66cadad8eed2811a2/CellDetection-0.4.8-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "3e4d110e19ea629c1e1702d0f3a65d977dabcb2e5d9c652658961c64fe99955d",
"md5": "fe4e5da85b6f210de920d7dfcc3838ef",
"sha256": "efa8fbb83c168c26b8e5b2d949d6df8feb104a06509531e36a801ee48408a331"
},
"downloads": -1,
"filename": "celldetection-0.4.8.tar.gz",
"has_sig": false,
"md5_digest": "fe4e5da85b6f210de920d7dfcc3838ef",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 153736,
"upload_time": "2024-05-29T14:36:59",
"upload_time_iso_8601": "2024-05-29T14:36:59.986985Z",
"url": "https://files.pythonhosted.org/packages/3e/4d/110e19ea629c1e1702d0f3a65d977dabcb2e5d9c652658961c64fe99955d/celldetection-0.4.8.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-05-29 14:36:59",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "celldetection"
}
JSON