Title :
A Codeword Visualization Tool for Dense Trajectory Feature
Author :
Phan, Sang ; Lam, Vu ; Tran, Son ; Ngo, Thanh Duc ; Le, Duy-Dinh ; Satoh, Shin Ichi
Author_Institution :
Univ. of Sci., Ho Chi Minh City, Vietnam
Abstract :
Bag of Words model is one of the most popular approaches to various classification tasks. Visualizing code words can help us deeply understand the discriminative power of each feature in the model. In practice, it is difficult to visualize code words, especially in case of motion feature because the codebook size is often very large. In this demo, we introduce a codeword visualization tool for the dense trajectory motion feature. Our visualization software is based on user´s interest points, where the end user can select a specific codeword and see how that codeword appears in the video dataset.
Keywords :
data visualisation; image classification; image motion analysis; video signal processing; bag of words model; codebook size; codeword visualization tool; dense trajectory motion feature; users interest points; video dataset; visualization software; Computational modeling; Conferences; Data visualization; Head; Tracking; Trajectory; Visualization; bag-of-words; trajectory; visualization;
Conference_Titel :
Multimedia and Expo Workshops (ICMEW), 2012 IEEE International Conference on
Conference_Location :
Melbourne, VIC
Print_ISBN :
978-1-4673-2027-6
DOI :
10.1109/ICMEW.2012.128