The following repository contains a web app for users to label their raw image data, expediting the process of data cleaning and labelling.
Users may upload a zip file where the app will then sort and return the images in folders titled with the predicted image class based on the model utilized.
Using your own pretrained model is as simple as importing it and changing a single line of code in the app.py file.
- Clone/download this repository
- Install dependencies
pip install -r requirements.txt
- Launch the web app
python app.py
- Open your browser and head over to localhost:8000/landing_page and begin using the app
Alternatively, try it out with Docker:
docker run -d -p 8000:8000 danlimwz/flask_image_classifier
The classification model uses a ResNet50 architecture which was pretrained on the ImageNet dataset. Visit this link for supported image classes: http://image-net.org/explore
The box and mask model uses Mask RCNN trained on the COCO dataset. Visit this link to explore the classes supported: http://cocodataset.org/#explore
Check out the article I wrote on medium for a tutorial on how to create the app: https://medium.com/@limwz.daniel/deploying-your-deep-learning-model-using-flask-and-docker-c05a6d1d96a5
- Image class selection
- Confidence threshold selection
- Facial recognition