2018/08/22 by Tsung-Yu Lin, Tsung‐Yu Lin, Lin, Tsung-Yu +4 · 3 citations
Computer Science · #Advanced Image and Video Retrieval Techniques #Advanced Neural Network Applications #Algorithm #Artificial intelligence #Computer Vision and Pattern Recognition (cs.CV) #Computer science #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Machine learning #Normalization (sociology) #Pattern recognition (psychology) #Pooling #Theoretical computer science #cs.CV
paper · pdf · doi:10.48550/arxiv.1808.07503
published in arXiv (Cornell University) (Cornell University)
arxiv created 2018/08/22 · openalex publication_date 2018/08/22 · arxiv updated 2018/08/24 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Aggregated second-order features extracted from deep convolutional networks have been shown to be effective for texture generation, fine-grained recognition, material classification, and scene understanding. In this paper, we study a class of orderless aggregation functions designed to minimize interference or equalize contributions in the context of second-order features and we show that they can be computed just as efficiently as their first-order counterparts and they have favorable properties over aggregation by summation. Another line of work has shown that matrix power normalization after aggregation can significantly improve the generalization of second-order representations. We show that matrix power normalization implicitly equalizes contributions during aggregation thus establishing a connection between matrix normalization techniques and prior work on minimizing interference. Based on the analysis we present γ-democratic aggregators that interpolate between sum (γ=1) and democratic pooling (γ=0) outperforming both on several classification tasks. Moreover, unlike power normalization, the γ-democratic aggregations can be computed in a low dimensional space by sketching that allows the use of very high-dimensional second-order features. This results in a state-of-the-art performance on several datasets.