Unsupervised Learning of Object Features from Video Sequences
- 27 July 2005
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
Abstract
We develop an efficient algorithm for unsupervised learning of object models as constellations of features, from low resolution video sequences. The input images typically contain single or multiple objects that change in pose, scale and degree of occlusion. Also, the objects can move significantly between consecutive frames. The content of an input sequence is unlabeled so the learner has to cluster the data based on the data's implicit coherence over time and space. Our approach takes advantage of the dependent pairwise co-occurrences of objects' features within local neighborhoods vs. the independent behavior of unrelated features. We couple or decouple pairs of features based on a probabilistic interpretation of their pairwise statistics and then extract objects as connected components of features.Keywords
This publication has 6 references indexed in Scilit:
- Segmenting, modeling, and matching video clips containing multiple moving objectsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2004
- Tractable group detection on large link data setsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2004
- Object class recognition by unsupervised scale-invariant learningPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2003
- Using temporal coherence to build models of animalsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2003
- Affine-invariant local descriptors and neighborhood statistics for texture recognitionPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2003
- Object recognition from local scale-invariant featuresPublished by Institute of Electrical and Electronics Engineers (IEEE) ,1999