High-Level Feature Extraction Experiments for TRECVID 2007

1. Briefly, what approach or combination of approaches did you test in each of your submitted runs? A_KL1_1: A color-based image retrieval method using three kinds of image features: a global color distribution feature, a common bitmap feature and a Wavelet texture feature. Key-frames generated by our frame clustering method with threshold 5 were used as the input of the feature extraction system. A_KL2_2: A color-based image retrieval method in the same way as A_KL1_1, where key-frames generated with threshold 20 were used as the input. A_KL3_3: SVMs based on three visual features: a modified MPEG-7-based edge histogram descriptor, a color layout descriptor and an auto-correlogram, where key-frames generated with threshold 5 were used as the input data. A_KL4_4: SVMs as for A_KL3_3 and nine kinds of Haar-like feature-based extractors were used. A_KL5_5: In addition to A_KL4_4, a Haar-Like feature-based face extractor was applied to extract human related features. A_KL6_6: In the same way as A_KL5_5, but the HaarLike feature-based extractor with lower recall and higher precision was used.