Hi,
I would like to understand well the "Manifold-Ranking Based Image Retrieval" principale espcially using positive and negative samples.
Knowing that the images of the database are classified in different classes.
According to me, Manifold-Ranking Based Image Retrieval algorithm is as follows:
1.assign postive ranking score to the query (1) and zero to the remainng points of the Image database.
2.generate a weighted graph:
-Compute K nearest neighbors for each image
-Connect 2 images with edge if they are neighbors
-Form the affinity matrix to define the edge weights
-Normalise the affinity matrix
3. spread the score of all images to the neighbors via the weighted graph until the images (which have ranking score Zero ) get ranking score different from Zero.
4. rank the images according to their ranking scores (largest ones in the top)
Now, I present Manifold-Ranking Based Image Retrieval using relevance feedback (positive and negative images) algorithm
5. Active leanrning selects the positive and negative images.
6. Rerun Manifold-Ranking Based Image Retrieval algorithm
Q.1 The stated algorithms are correct?
Q.2 When I execute different queries for the same class, I get the same retrieved images but with different ranking score for each query . Could you explain why?
Q.3 How to compute the effectiveness of the image retrieval?
Attached, the concerned paper.
Thank you a lot.
Best Regards.
Khadidja Belattar.