How to find image retrieval using SIFT ?

I'm doing research in image retrieval using SIFT features. how to compute SIFT features for images and compare with the SIFT feature of query image using euclidean distance and retrieve the top 10 best match images alone?

Please Send me a Code.....

Answer 1

[Gautam]

Hello jenifer,

You can use the “detectSIFTFeatures” and “extractFeatures” functions from the Computer Vision Toolbox to extract SIFT features in MATLAB.

Then, to compare the SIFT features with those of the query image using Euclidean distance, and rank them based on similarity, use the “pdist2” function.

distances = pdist2(queryFeatures, databaseFeatures, 'euclidean');

Here are the links to these functions which you can refer for more information:

1. detectSIFTFeatures : https://www.mathworks.com/help/vision/ref/detectsiftfeatures.html
2. extractFeatures: https://www.mathworks.com/help/vision/ref/extractfeatures.html
3. pdist2: https://www.mathworks.com/help/stats/pdist2.html