Image comparison

The task is to develop an application that compares a pair of images and computes a similarity score (i.e., how similar the two images are on a 0 to 1 scale). You will be using an existing image matching algorithm called Scale-Invariant Feature Transform (SIFT). If you are not familiar with SIFT, you can start with the Wikipedia page of the algorithm: [login to view URL] Towards the bottom of the page you will find links to a variety of information on SIFT including this pointer to an open-source SIFT implementation in C/C++: [login to view URL] So, you may not need to code the algorithm yourself; you will only need to build an application to use the algorithm for comparing the images that I provide. What I will provide you is a folder containing several text files--each containing URLs to some images (80 on average). The images will be in a variety of formats, e.g., jpg, png, gif, tif, and psd. Your job will be to compute the similarity scores of every pair of image within each file using the SIFT algorithm. Example: Suppose the folder that I give you contains three text files T1, T2, and T3, and suppose T1 contains URLs to 5 pictures--P1, P2, P3, P4, and P5. Then your job will be to use the SIFT algorithm to compute the similarity scores of the image pairs P1-P2, P1-P3, P1-P4, P1-P5, P2-P3, P2-P4, P2-P5, P3-P4, P3-P5, and P4-P5. You will have to do the same for images contained in the files T2, and T3. You can write the output (i.e., the scores) corresponding to each file in a spreadsheet. We have hundreds of such text files but for the development and testing of the application, we will put 10 to 20 files in a folder. Once you are done with your development and testing, we will run the program on all the files that we have. So, you have to code the application in such a way that it goes through all the text files in a certain folder. To summarize, the deliverables should be (i)an easy to run application that we can use to compute the pair-wise scores for the images in all the text files and write them on a spreadsheet, as explained above, and (ii) the application code that we can modify later if needed.