This is an object detection flask web app running with tenssorflow server. It can classify 90 different classes of common objects. The algorithm Faster-RCNN has been trained using TensorFlow Object Detection API.
This file contains instructions on executing the commands to run object detection web app. The steps are given below.
- Command to build docker file for the object detection web app with the name detector-app.
sudo docker build -t detector-app .
- Command to create network binding and run Tensorflow server with GPU support.
sudo docker run --gpus all -p 8500:8500 --name detector --mount type=bind,source=/home/msc1/Desktop/Labs/Semester_2/7147COMP/coursework/saved_model,target=/models/detector -e MODEL_NAME=detector -t tensorflow/serving:latest-gpu
- Command to run web app directly, open another terminal and go to FlaskObjectDetection directory. Run the following command.
python app.py
- Command to run the web app docker image created in step 1.
sudo docker run -p 127.0.0.1:5000:5000 detector-app
Some screenshots from the webapp is given below.
