A Method for Mobile Photo Recommendation
Seok Hoon Kang · Advanced materials research · 2013
Spending a lot of time to find right photos from a large album in a mobile device is very inefficient. To solve the problem, various methods have been suggested, which automatically present the photos that users are looking for. In this paper, in order to automatically choose photos that a user wants, a photo recommendation system using visual information is proposed. The recommendation method divides photos and detects visual feature points using entropy. From the experiment of using the photo recommendation program suggested in this paper, the users showed more than 90% of satisfaction rate. As the result of searching for a specific photo, the suggested method by this paper has the higher F-measure by average 17% than the existing recommendation methods. Furthermore, its searching time is 4 times faster than that of existing methods