## GENERATING REPRESENTATIVE SETS AND SUMMARIES FOR LARGE COLLECTION OF IMAGES: EVALUATION (2)

As we have 3 sets of each image set which contains 10, 15 and 20 images respectively, we follow the fixed pattern for each set which can be seen in figure 6.

Figure 6: Questionnaire pattern for each result sets

To understand deeply, we would like to give an example here. Suppose we are evaluating one of the representative set called “I3-75”. As we mentioned earlier that we consider 20 images in a representative set, first we ask to rate 10 top ranked images generated from ranking mechanism and then we show 15 top ranked images and ask participants whether they think that these 15 images are more or less representative then 10 images. If they say yes, then we ask for the rating again otherwise we directly show whole 20 images summary and asking the same question in perspective of 10 and 15 images summary. Again if they say “yes”, then we ask again rating for the 20 images summary. The pattern of the questionnaire will be same for rest of the five result sets.

Figure 7 : Time consumed by different cropped window

From the above graph, there is nominal difference in the time taken by windows generating algorithm (blue bar) for N=3 and N=5.The time taken by clustering algorithm is higher for both N=5, C=85% random and sequence windows. The reason of this is to fetch sift descriptors of all the windows and then process for the clustering. While for the set N=3, C=66% sequence took very less time. There is no much difference in time taken by the ranking mechanism for all sets. The set N=5, C=75% sequence windows took long time for the ranking mechanism in compare to others. Maximum total time is consumed by N=5, C=85% random windows set. The time taken by clustering and fetching centroids increases with the increase of number of windows N and the coverage C.