This paper describes our participation in the NIST TERCVID 2004 retrieval evaluation. In the first-year effort for the TERCVID project, we only tackle Person X detection of the high-level feature extraction task. We design an automatic Person X detector using frontal faces in videos solely. We illustrate the architecture of Person X detector and the evaluation results in this paper.

We present in this paper a study on auditory feature spaces for speech-driven face animation. The goal is to provide solid analytic ground to underscore the description capability of some well-known features with relation to lipsync. A set of various audio features describing the temporal and spectral shape of speech signal has been computed on annotated audio extracts. The dimension of the inp...

We describe our fourth participation, that includes two high-level feature extraction runs, and one manual search run, to the TRECVID video retrieval evaluation. All of these runs have used a system trained on the common development collection. Only visual information, consisting of color, texture and edge-based low-level features, was used.

in this paper, the performance of 11 different distances for image retrieval and classification, based on color, shape and texture, is evaluated. the precision-recall measure and the correct classification rate of the k-nn classifier are used to evaluate retrieval and classification performances, respectively. the experimental results for a database of 1000 images from 10 different semantic gro...

We studied two methods for the high-level feature extraction (HLF) task: (1) a method based on support vector machines (SVMs) with walk-based graph kernels [1], and (2) a method based on the prefixspan boosting (pboost) algorithm [2]. In the former method, each image is first segmented into a finite set of homogeneous segments and then represented as a segmentation graph where each vertex is a ...

 A_CL1_1: choose the best-performing classifier for each concept from all the following runs and an event detection method.  A_CL2_2: (visual-based) choose the best-performing visual-based classifier for each concept from runs A_CL4_4, A_CL5_5, and A_CL6_6.  A_CL3_3: (visual-text) weighted average fusion of visual-based classifier A_CL4_4 with a text classifier.  A_CL4_4: (visual-based) ave...

We present a new system for the recognition of cursive handwriting that is based on a perceptive model and neural networks. At the high level, our system takes into account several psychological effects such as the word superiority effect. At the low level, it utilizes a global feature extraction method which models how some features might be preattentively detected by the human visual system. ...

In this paper, the system developed by the University of Bremen for participation in the Trecvid 2006 high-level feature extraction task is presented. Six runs have been submitted, each of them incorporating a different combination of three classifiers based on image, sound, and text features. For the feature Corporate Leader, aboveaverage results could be achieved. Results are shown and differ...