This paper presents the systems used by CLIPS-IMAG laboratory. We participated to shot segmentation and high-level extraction tasks. We focus this year on High-Level Features Extraction task, based on key frames classification. We propose an original and promising framework for incorporating contextual information (from image content) into the concept detection process. The proposed method combines local and global classifiers with stacking, using SVM. We handle topologic and semantic contexts in concept detection performance and proposed solutions to handle the large amount of dimensions involved in classified data.