Dear Nisreen

You need to define each image by a feature vector, those features should be invariant to scale, rotation and translation, and you may perform the following steps:

1- Extract some features from the query image, and save it an array called QArray.

2- For each image in the database, extract the same features in step 1 in the same order and save the features in array called ImageArrayi.

3- For each ImageArrayi , calculate the similarity with QArray, using a similarity (distance) function

4- Define a threshold t,

5- For each i in ImageArrayi  if the distance from QArray to ImageArrayi < t, then retrieve image i, else do nothing,

====

What are the features?

There are hundreds of them, you may use:

1- color histogram (probabilities of the RGB colors in the image

2- you may use the intensity histogram, (probabilities of the (R+G+B)/3 in the image

3- Statistic features, such as mean, mode, media stdev, moments, uniformity, entropy, etc.

4- Hundreds more, read about them.

====

What are the similarity measures?

1- Euclidian distance

2- Manhattan distance

3- Hassanat Distance, this one is invented by me, and it is better than the previous ones because it is invariant to the outliers. you may find the link of the paper at: https://www.researchgate.net/publication/264995324_Dimensionality_Invariant_Similarity_Measure

4- tens of other distances..

Contributions in such research is highly dependent on the features used.

For your first move, you may start with the color features.

Article Dimensionality Invariant Similarity Measure

Dear Nisreen

You need to define each image by a feature vector, those features should be invariant to scale, rotation and translation, and you may perform the following steps:

1- Extract some features from the query image, and save it an array called QArray.

2- For each image in the database, extract the same features in step 1 in the same order and save the features in array called ImageArrayi.

3- For each ImageArrayi , calculate the similarity with QArray, using a similarity (distance) function

4- Define a threshold t,

5- For each i in ImageArrayi  if the distance from QArray to ImageArrayi < t, then retrieve image i, else do nothing,

====

What are the features?

There are hundreds of them, you may use:

1- color histogram (probabilities of the RGB colors in the image

2- you may use the intensity histogram, (probabilities of the (R+G+B)/3 in the image

3- Statistic features, such as mean, mode, media stdev, moments, uniformity, entropy, etc.

4- Hundreds more, read about them.

====

What are the similarity measures?

1- Euclidian distance

2- Manhattan distance

3- Hassanat Distance, this one is invented by me, and it is better than the previous ones because it is invariant to the outliers. you may find the link of the paper at: https://www.researchgate.net/publication/264995324_Dimensionality_Invariant_Similarity_Measure

4- tens of other distances..

Contributions in such research is highly dependent on the features used.

For your first move, you may start with the color features.

Article Dimensionality Invariant Similarity Measure

Dear Nisreen,

First of all, I completely agree with the response of Ahmad, the key in any CBIR problem is to implement a robust description of the images by computing effective visual features. Besides, the subsequent matching of these features with the aim of obtaining the similarity between the images is very important before the implementation of the indexed image database.

My research line is not focused on medical images, but I have worked in other kind of CBIR problems, such as place recognition. Related to this, in the last year we have developed a method called ABLE (Able for Binary-appearance Loop-closure Evaluation). Although this algorithm is focused on place recognition, it could be adapted to your CBIR problem. We use LDB features and a matching based on the Hamming distance which is very fast and effective. I attach you some publications where the method is described.

I hope that my answer helps you. Good luck with your research!

Conference Paper Fast and Effective Visual Place Recognition using Binary Cod...

Conference Paper Bidirectional Loop Closure Detection on Panoramas for Visual...

Thank you,

Original images are gray scale image, and I have segmented whats called "Cerebral Sacular Aneurysms(CSA)" from gray scale images and the results are binary images .

I have used shape features in segmentation stage to extract CSA Now I want:

1- to build an indexed image database, depending on the shape features.

2- Apply a retrieval engine to retrieve images that match a query image.

** I am using MATLAB as a platform,

What is a real help for me is an example about any CBIR??

Thank you Mr. Mathew,

1-The purpose of the CBIR is to retrieve images from an indexed image database that have a "similar"  cerebral sacular aneurysms. Matching should relate to size"area", orientation, Minor axes, major axes ...

2- What is preferable "acceptable" size of the segmented database? I have about 20 segmented images.

3- Query image is a Digital Subtraction Angiography image, User can choose the image from a specified path.

My advice ,you can extract set of features of database offline and save it in excel sheet as a matric two dimension each row set of features for an image and then apply ecludian or any other techniques to check whether input image is similar to the image in database or not...

also yiu can depend on edge to extract the features

good luck 

Dear Nisreen

You can see the link

www.ijsce.org/attachments/File/v3i1/A1311033113.pdf

Regards

prof. Dr. Muzhir Al-Ani

There has been intenstive research in the field of CBIR and many CBIR systems are now in existence. Google similar image search is a typical example. 

However, a note of caution. The biggest challenge in CBIR is the "semantic gap" between the visual content features and the semantics of the image itself. Despite good research results, practical CBIR systems performance is still unsatisfactory. 

What you are working on is special medical images which have a lot of domain specific features where the general solutions would nor work well. We are also working on medical images (X-ray, ultrasouund, CT scan), and have encountered quite a lot of problems. It seems that a large amount of domain knowledge in the application area is required to be successful.

Please see some work reported on my research gate page.

Good luck.

Dear Nisreen

I am working a thesis on this area and now i am looking for a textbook related with image processing, Therefore can you show me the link which can i go through since your title is very similar with me and i can't find any reference textbook which talks more about image processing with Matlab.

best wishes, Behayilu

http://www.ijsce.org/attachments/File/v3i1/A1311033113.pdf

if you are working with matlap the best source matlab library . This source maybe will work.