Compact bilinear pooling via kernelized random projection for fine-grained image categorization on low computational power devices
| Autor: | Juan M. Corchado, Angélica González Arrieta, Daniel López-Sánchez |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
0209 industrial biotechnology
Computer science Cognitive Neuroscience Random projection Pooling Bilinear pooling Bilinear interpolation 02 engineering and technology Polynomial kernel Image (mathematics) 020901 industrial engineering & automation Dimension (vector space) Artificial Intelligence 0202 electrical engineering electronic engineering information engineering 1203.17 Informática business.industry Deep learning Computer Science Applications 1203.04 Inteligencia Artificial 020201 artificial intelligence & image processing Artificial intelligence business Algorithm |
| Zdroj: | GREDOS: Repositorio Institucional de la Universidad de Salamanca Universidad de Salamanca (USAL) GREDOS. Repositorio Institucional de la Universidad de Salamanca instname |
| ISSN: | 0925-2312 |
| DOI: | 10.1016/j.neucom.2019.05.104 |
| Popis: | Bilinear pooling is one of the most popular and effective methods for fine-grained image recognition. However, a major drawback of Bilinear pooling is the dimensionality of the resulting descriptors, which typically consist of several hundred thousand features. Even when generating the descriptor is tractable, its dimension makes any subsequent operations impractical and often results in huge computational and storage costs. We introduce a novel method to efficiently reduce the dimension of bilinear pooling descriptors by performing a Random Projection. Conveniently, this is achieved without ever computing the high-dimensional descriptor explicitly. Our experimental results show that our method outperforms existing compact bilinear pooling algorithms in most cases, while running faster on low computational power devices, where efficient extensions of bilinear pooling are most useful. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect