API created using Flask. It will allow the user to send an url of a dog image and will return the number and positions of the dogs within the image. Built with imageai.
Download the weights for the neural network here. And place them in this project folder. Run the docker compose file to install all the dependencies and create the build.
docker-compose up --build -d
It will install all the dependencies for python and start the service in the port 80. With the -dflag we will indicate it to run in the background.
Send the url of a dog image in a json request with the method POST. It will return the number an array of the dogs within the image.
It will have to be requested in the /predictroute.
Using the following image:
We will just send the following json to our api (using the url of the image)
And will return us the response with the dogs within the image.
An example for several images of dogs can be seen in a Jupyter notebook in the folder jupyterExample. As well as the answers to the questions asked.
Implemented with the detection library imageai.