Feature Extraction and Reverse Image Search on Caltech-101 Dataset

This notebook will guide you through the process of analyzing an image dataset using a pre-trained convolution network (VGG16) and extracting feature vectors for each image

Post analysis we try to demonstrate 'reverse image search' one of the widely popular applications of image analysis.

Flow:

• Download VGG16 pre-trained model using keras

• Perform Feature Extraction :

  Here we remove the last layer ie.,the softmax classification layer so our output model now has only 12 layers and the last layer would be fc2(Dense) a fully connected layer

• Get feature vectors for all the images then scale them down using PCA

• Use cosine distance between pca features to compare the query image to 5 number of closest images and return them as thumbnails

Watch the Video:

Steps:

Deploy on IBM Cloud:

• Sign up for IBM's Watson Studio. By creating a project in Watson Studio a free tier Object Storage service will be created in your IBM Cloud account. Take note of your service names as you will need to select them in the following steps.

Note: When creating your Object Storage service, select the Free storage type in order to avoid having to pay an upgrade fee.

• Create a new Project in Watson Studio (New --> Standard project)
• Create an new Notebook (Add to project --> Notebook --> from url)
• Create an Environment (Environment --> New Environment -->)
• Provision the notebook on newly created Environment
• Run the project

• Stop the Environment Post usage

Run Locally:

• Clone the repository

git clone https://github.com/rapchic/Reverse-Image-Search.git

• Navigate into the directory

cd Reverse-Image-Search-master

• Run using Jupyter notebooks

jupyter notebook

Sample Output:

License

This code pattern is licensed under the Apache License, Version 2. Separate third-party code objects invoked within this code pattern are licensed by their respective providers pursuant to their own separate licenses. Contributions are subject to the Developer Certificate of Origin, Version 1.1 and the Apache License, Version 2.

Apache License FAQ