This repository includes code for receiving frontend data(images, bounding boxes) from rabbitmq, creates dataset for the new image in PASCAL VOC format, and initiates training(fine-tuning) on the new and old images in py-faster-rcnn.
Prerequisites:
Install Rabbitmq
composer.json is in the repo. Execute "composer.phar install"
The above command is used to install "The php-amqplib client library". For any trouble please follow the instructions in "https://www.rabbitmq.com/tutorials/tutorial-one-php.html"
Project Structure:
009963.xml -> Sample xml format to save the image in PASCAL VOC format
detect_objects -> This file should be placed in py-faster-rcnn/tools
create_xml.py -> This file generates xml file for each image(example format 009963.xml)
image_processing.py -> Dummy code that returns bounding boxes
listdir.py -> returns the number of images in PASCAL VOC dataset. This is auxilary file for create_xml.py
receiver.py -> Communicates with rabbitmq and extract all the data sent from GUI.
server.py -> starts rabbitmq server.