Title :
Painting analysis using wavelets and probabilistic topic models
Author :
Tong Wu ; Polatkan, Gungor ; Steel, David ; Brown, Walter ; Daubechies, Ingrid ; Calderbank, R.
Author_Institution :
Dept. of Electr. & Comput. Eng., Rutgers Univ., Piscataway, NJ, USA
Abstract :
In this paper, computer-based techniques for stylistic analysis of paintings are applied to the five panels of the 14th century Peruzzi Altarpiece by Giotto di Bondone. Features are extracted by combining a dual-tree complex wavelet transform with a hidden Markov tree (HMT) model. Hierarchical clustering is used to identify stylistic keywords in image patches, and keyword frequencies are calculated for sub-images that each contains many patches. A generative hierarchical Bayesian model learns stylistic patterns of keywords; these patterns are then used to characterize the styles of the sub-images; this in turn, permits to discriminate between paintings. Results suggest that such unsupervised probabilistic topic models can be useful to distill characteristic elements of style.
Keywords :
Bayes methods; feature extraction; hidden Markov models; pattern clustering; probability; trees (mathematics); unsupervised learning; wavelet transforms; Giotto di Bondone; HMT model; Peruzzi Altarpiece; computer-based techniques; dual-tree complex wavelet transform; feature extraction; generative hierarchical Bayesian model; hidden Markov tree model; hierarchical clustering; image patches; keyword frequencies; keyword stylistic pattern learning; probabilistic topic models; stylistic keyword identification; stylistic painting analysis; stylistic pattern learning; subimage style characterization; unsupervised probabilistic topic models; Hidden Markov Trees; Machine Learning; Painting Analysis; Topic Models; Wavelet Transforms;
Conference_Titel :
Image Processing (ICIP), 2013 20th IEEE International Conference on
Conference_Location :
Melbourne, VIC
DOI :
10.1109/ICIP.2013.6738672
