https://github.com/nkottary/pca-image-search

Content Based Image Retrieval Using Latent Semantic Indexing.
https://github.com/nkottary/pca-image-search

Last synced: 27 days ago
JSON representation

Content Based Image Retrieval Using Latent Semantic Indexing.

• Host: GitHub
• URL: https://github.com/nkottary/pca-image-search
• Owner: nkottary
• Created: 2014-08-23T16:27:29.000Z (about 10 years ago)
• Default Branch: master
• Last Pushed: 2014-08-23T16:38:47.000Z (about 10 years ago)
• Last Synced: 2024-10-03T05:09:45.760Z (about 1 month ago)
• Language: Matlab
• Size: 133 KB
• Stars: 0
• Watchers: 2
• Forks: 6
• Open Issues: 0
• Metadata Files:
  • Readme: Readme.txt

Awesome Lists containing this project

README

        
Content Based Image Retrieval Using Latent Semantic Indexing

------------------------------------------------------------

Authors:

--------

1) Venkatesh D. 		10IT102		[email protected]		

2) Sambhrum I. G. 		10IT84		[email protected]	

3) Pradeep P. 			10IT60		[email protected]

4) Nishanth H. Kottary 		10IT54		[email protected]	

Requirements:

-------------

Matlab R2011a.

Directories:

------------

/report/			Contains report details.

	/latex/			Contains latex files and images.

	/pdf/			Contains the latex files compiled to pdf.

/images/			Folder containing 679 images.

/vectors/			Folder containing all the extracted features

	/reduction/		Folder containing reduced vectors.

	/clustering/		Folder containing clustered vectors. 

/matlab code/			Contains all the code files.

	/clustering/		Contains code for clustering.

	/reduction/		Contains code for reduction techniques.

	/extractors/		Contains code for extracting features.

	/scripts/		Contains code for automated tasks.

	/similarityMatchers 	Contains code for similarity metrics.

Instructions:

-------------

1) Generating Vectors:

- in /images/ folder place the image dataset. (you can download images from "INRIA Holidays Dataset" link: "http://lear.inrialpes.fr/je-gou/data.php".)

- Open the directory /matlab code/scripts/

- Run SaveFeatures.m to extract feature vectors.

- Run SavePCAVectors.m, SaveDualPCAVectors.m and SaveKPCAFeatures.m to generate the reduced vectors.

- Run SaveClusters.m to perform clustering for all the above vectors.

2) Running Application.

- Run /matlab code/GUI.m, this opens GUI application.

- Select similarity metric from drop down menu.

- Select whether to use threshold or not.

- Select whether to use reduced vectors or not.

- If reduced vectors are used then select one among PCA, D-PCA or K-PCA.

- Select whether to use clustering.

- Select number of clusters to look up from drop down box.

- Select the image to use as query.

- Click on retrieve button to get the results.

Frequently Asked Questions:

---------------------------

1) How to change the number of dimensions during reduction ?

- Open the reduction script (SavePCAVectors.m, SaveDualPCAVectors.m and SaveKPCAFeatures.m) and change the variable "dimension".

- Re run the scripts.

2) How to change the order of the kernel in K-PCA ?

- Change the variable "para" in the scripts SaveKPCAFeatures.m and kpcaReduction.m.

- Run the script SaveKPCAFeatures.m

3) Querying with image from external directory gives an error.

- This may be because the query image is not in path, in this case add the image to path.

- The image has to have a numerical filename. For Example "123.jpeg".

4) K-PCA gives me very bad results.

- This is because the "para" variable value in /matlab code/scripts/SaveKPCAFeatures.m is different from the value in /matlab_code/reduction/kpcaReduction.m, they need to be the same.

THANK YOU