Talk 4.4: 1:30PM–2:45PM
| Session Name: | Signal Processing and Network Applications |
| Session Time: | 1:30PM–2:45PM |
| Author Name: | Alexandre Karpenko |
| Author Email: | alex.karpenko@utoronto.ca |
| Talk Title: | Content-based Image Search using the Scale Invariant Feature Transform |
| Slides: | 4-4.ppt |
| Abstract: | In this paper we present a method for automatically finding common points of interest, known as tags, across images. Content-based image searches can then be performed using these automatically assigned tags. Better results are obtained than with searchable image repositories—such as Flickr—which use manual tagging. The Scale Invariant Feature Transform (SIFT) is used to find images that have certain keypoints in common. SIFT keypoints have been shown to be reliably identifiable under varying scales and levels of illumination. These properties have made them particularly suitable for applications such as object detection, panorama stitching, and camera tracking. In our specific work auto-tagging is performed by finding similar keypoints, comparing their nearest neighbours, and checking for keypoint localization. Our system—developed by the Artificial Perception Laboratory at the University of Toronto—is capable of fast and reliable automatic image tagging using this method. |
| Research Group: | Computer |
| Degree Program: | M.A.Sc. |
| Author Bio: | Alexandre Karpenko is a graduate student at the University of Toronto. He received a BASc in the Engineering Science Computer Option from the University of Toronto. He is currently pursuing a MASc in computer engineering in the area of machine learning and computer vision. |