A paper accepted to CVPR 2015

Our CVPR 2015 paper introduces novel visual similarity networks with applications in unsupervised image alignment.

The TUT capital gain funded project "BigData" resulted to a novel concept of "visual similarity graphs" which was developed in the Vision Group of the Dept. of Signal Processing. The method automatically constructs a graph (network) where similar images have strong links between them and the network represents visual similarities in natural way allowing efficient graph algorithms in large scale visual processing. The results will be presented in the IEEE Conf. on Computer Vision and Pattern Recognition (CVPR) in Boston, Massachusetts, USA in June 2015 in

  • Unsupervised Visual Alignment with Similarity Graphs (F. Shokrollahi Yancheshmeh, K. Chen, J.-K. Kämäräinen), In IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 2015.
CVPR is the premier annual Computer Vision event. According to Google Scholar Metrics, CVPR is the top publication venue in the field of computer vision and pattern recognition. CVPR also ranks 7th in the category of Engineering & Computer Science, and it is the highest-ranked venue in Computer Science.
Visual similarity graph of motorbikes
Visual similarity graph of motorbikes

Back to news...