https://github.com/i008/aiscope
malaria detection using retinanet object detection
https://github.com/i008/aiscope
deep-learning malaria object-detection retina-net
Last synced: 7 months ago
JSON representation
malaria detection using retinanet object detection
- Host: GitHub
- URL: https://github.com/i008/aiscope
- Owner: i008
- License: mit
- Created: 2018-04-04T22:38:42.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2018-06-05T21:01:07.000Z (over 7 years ago)
- Last Synced: 2025-01-12T14:38:00.487Z (9 months ago)
- Topics: deep-learning, malaria, object-detection, retina-net
- Language: Jupyter Notebook
- Homepage:
- Size: 19.8 MB
- Stars: 0
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.MD
- Changelog: HISTORY.rst
- Contributing: CONTRIBUTING.rst
- License: LICENSE
- Authors: AUTHORS.rst
Awesome Lists containing this project
README
### AiScope
http://aiscope.net/
------------------
Detecting malaria parasites in thick blood samples using retinanet object detection architecture
https://arxiv.org/abs/1708.02002
### Setup
This package requires
- https://github.com/i008/detdata
- https://github.com/i008/keras-retinanet
Install in develop mode:
```bash
python setup.py develop
```
### Inference
Check the ./notebooks directory for examples how to predict and train the detector
### Example (inference)
### Augmentation
Augmentation is done in detdata using the amazing imgaug package (https://github.com/aleju/imgaug)
### Run the demo (docker required)
To test on other images - place them in the testimages folder.
```bash
git clone https://github.com/i008/aiscope.git && cd aiscope
# run the container
sudo docker run -p 7777:7777 \
-v $PWD/testimages:/opt/aiscope/images \
-t i008/aiscope:latest jupyter notebook \
--no-browser \
--port 7777 \
--allow-root \
--ip=0.0.0.0 \
--NotebookApp.token=''
#go to localhost:7777
#open, run, play with demo.ipynb
```
### Results
