Gradiant at the International Conference on Computer Vision (ICCV 2011)

From November 6th to 13th, the International Conference on Computer Vision 2011 (ICCV’2011) was held in Barcelona (Spain). Long Long Yu and Pablo Dago Casas attended this conference, where the most important advances in computer vision were presented, as improvements in classification algorithms, human behaviour analysis, object recognition or the numerous applications of Kinect.

Within the ICCV, some tutorials about different subjects took place, among which we would like to stress two of them due to relation to some of the projects in which Gradiant is involved: “Decision forests for classification, regression, clustering and density estimation” (A. Criminisi and J. Shotton, Microsoft Research Cambridge) and “Geometry constrained parts based detection” (S. Lucey and J. Saragih, CSIRO, Australia). The main conference consisted of different sessions, each of them addressing different computer vision problems such as “Detection and categorization”, “Illumination and reflectance”, “Faces”, “Scene understanding”, etc.

Several workshops were organized jointly with the ICCV’2011, dealing with different aspects of computer vision as mobile vision, GPU processing or human-computer interfaces, in which Kinect played the leading role. At the workshop entitled “Benchmarking Facial Image Analysis Technologies”, Gradiant in collaboration with the University of Vigo presented the paper “Single- and Cross- Database Benchmarks for Gender Classification under Unconstrained Settings” (work done by Pablo Dago Casas, Daniel González Jiménez, Long Long Yu and José Luis Alba Castro), in which a new standardized protocol for gender classification experiments is proposed in order to allow different authors to evaluate and fairly compare their results in this subject. In addition, this workshop included talks from Jonathon Phillips (NIST -National Institute of Standards and Technology-) and Maja Pantic (Imperial College London), both well known experts in different fields of facial image processing (biometrics and facial expressions respectively).

With its presence at this conference, Gradiant continues publishing its advances in facial image processing, and opened possible collaborations with international institutions as the research group Facial Image Processing and Analysis (FIPA, Karlsruhe Institute of Technology).