Prof. Ricardo Torres of RECOD, together with Prof. Daniel Pedronette (RECOD alumni, now at UNESP/Rio Claro) was among the nine finalists for the best paper award at this year’s IEEE International Conference on Image Processing (ICIP’2014) held last month in Paris, France, for the work “Unsupervised Manifold Learning by Correlation Graph and Strongly Connected Components for Image Retrieval”. Here’s the abstract :
This paper presents a novel manifold learning approach that takes into account the intrinsic dataset geometry. The dataset structure is modeled in terms of a Correlation Graph and analyzed using Strongly Connected Components (SCCs). The proposed manifold learning approach defines a more effective distance among images, used to improve the effectiveness of image retrieval systems. Several experiments were conducted for different image retrieval tasks involving shape, color, and texture descriptors. The proposed approach yields better results in terms of effectiveness than various methods recently proposed in the literature.
A link for the paper in the IEEE eXplorer will be soon available. The last preprint of the fulltext is already available.