R-aryan / Cards_Detection_Using_FASTER-RCNN

End to end object detection project using Faster RCNN, trained using TFOD1.14 framework and served as REST API via Flask App

Cards Detection Using FASTER RCNN

• End to end object detection project using Faster RCNN.
• The training is done using TFOD1.14(Tensorflow object detection) framework.
• The application is served as an REST API using Flask.
• Faster RCNN paper can be found here.

Steps to run the application

Setting up virtual environment.

• What is Virtual Environment in python ?
• Create virtual environment in python
• Create virtual environment Anaconda
• create a virtual environment and install requirements.txt

pip install -r requirements.txt

• After setting up the virtual environment download the trained weights from here.
• After downloading the trained weights place it under the directory services/card_detector/application/ai/weights/exported_inference_graph/
• After performing the above steps go to services/card_detector/api and run app.py

python app.py

• After running the app.py the web app can be accessed at http://127.0.0.1:9000/ copy this url and paste it in your browser.
• The UI will look like the following.

• The picture can be uploaded using the upload button and after uploading the image click on predict to perform inference.
• Sample Input

• Sample Output

• The application logs can also be found here.