Web Application with Django for Object Identification deployed in AWS Lightsail:

To try this web application click here

In this repo, I use Django to create a web application and I use TensorFlow as the backend to do object identification. I utilize Keras's ResNet50(pre-trained on ImageNet) to classify the given image. I also find ten most similar images among COCO-val2014 images to the given image. To achieve this, I load pretrained ResNet50 without fully connected layers and use it as a feature extractor. I extract features from COCO-val2014 images using the feature extractor and store these features in numpy arrays. For the given image, I also extract its features using pretrained ResNet50, and then by running the K-nearest neighbor algorithm in Euclidean space, I find the most similar images to the given image.

Project Requirements:

The file requirements.txt contains all required packages to run the project.

How to run:

git clone https://github.com/mshadloo/Web-Application-with-Django-for-Object-Identification.git
cd Web-Application-with-Django-for-Object-Identification
chmod +x run.sh && ./run.sh