Output details

A Robust and Scalable Visual Category and Action Recognition System using Kernel Discriminant Analysis with Spectral Regression

<24>Visual concept detection (VCD) and action recognition (AR) are the most important tasks in content-based multimedia information retrieval. They aim to annotate images using a vocabulary defined by a set of concepts of interest including scenes types (mountains, snow) or human actions (phoning, playing instrument). ImageCLEF PhotoAnnotation and Pascal Visual Object Classes challenges are state-of-the art evaluation initiatives intended to evaluate image-based approaches for VCD and AR. This paper describes an original solution for these challenges resulting in a system which was evaluated independently by challenge organisers and ranked first. This work was partially supported from EPSRC project ACASVA (EP/F069421/1)