We propose a novel topic discovery algorithm for unlabeled images based on the bag-of-words (BoW) framework. We first extract a dictionary of visual words and subsequently for each image compute a visual word occurrence histogram. We view these histograms as rows of a large matrix from which we extract sparse principal components (PCs). Each PC identifies a sparse combination of visual words which co-occur frequently in some images but seldom appear in others. Each sparse PC corresponds to a topic, and images whose interference with the PC is high belong to that topic, revealing the common parts possessed by the images. We propose to solve the associated sparse PCA problems using an Alternating Maximization (AM) method, which we modify for the purpose of efficiently extracting multiple PCs in a deflation scheme. Our approach attacks the maximization problem in SPCA directly and is scalable to high-dimensional data. Experiments on automatic topic discovery and category prediction demonstrate encouraging performance of our approach. Our SPCA solver is publicly available.
|Original language||English (US)|
|Title of host publication||Brain-Inspired Intelligence and Visual Perception|
|Publisher||Springer International Publishing|
|Number of pages||24|
|State||Published - Feb 14 2019|
Bibliographical noteKAUST Repository Item: Exported on 2020-10-01
Acknowledgements: This work was partially supported by the U.S. National Science Foundation, under award number NSF:CCF:1618717, NSF:CMMI:1663256 and NSF:CCF:1740796.