Fine-grained visual categorization (FGVC) has drawn increasing attention as an emerging research field in recent years. In contrast to generic-domain visual recognition, FGVC is characterized by high intra-class and subtle inter-class variations. To distinguish conceptually and visually similar categories, highly discriminative visual features must be extracted. Moreover, FGVC has highly specialized and task-specific nature. It is not always easy to obtain a sufficiently large-scale training dataset. Therefore, the key to success in practical FGVC systems is to efficiently exploit discriminative features from a limited number of training examples. In this paper, we propose an efficient two-step dimensionality compression method to derive compact middle-level part-based features. To do this, we compare both space-first and feature-first convolution schemes and investigate their effectiveness. Our approach is based on simple linear algebra and analytic solutions, and is highly scalable compared with the current one-vs-one or one-vs-all approach, making it possible to quickly train middle-level features from a number of pairwise part regions. We experimentally show the effectiveness of our method using the standard Caltech-Birds and Stanford-Cars datasets.
Hideki NAKAYAMA
University of Tokyo
Tomoya TSUDA
University of Tokyo
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Hideki NAKAYAMA, Tomoya TSUDA, "Efficient Two-Step Middle-Level Part Feature Extraction for Fine-Grained Visual Categorization" in IEICE TRANSACTIONS on Information,
vol. E99-D, no. 6, pp. 1626-1634, June 2016, doi: 10.1587/transinf.2015EDP7358.
Abstract: Fine-grained visual categorization (FGVC) has drawn increasing attention as an emerging research field in recent years. In contrast to generic-domain visual recognition, FGVC is characterized by high intra-class and subtle inter-class variations. To distinguish conceptually and visually similar categories, highly discriminative visual features must be extracted. Moreover, FGVC has highly specialized and task-specific nature. It is not always easy to obtain a sufficiently large-scale training dataset. Therefore, the key to success in practical FGVC systems is to efficiently exploit discriminative features from a limited number of training examples. In this paper, we propose an efficient two-step dimensionality compression method to derive compact middle-level part-based features. To do this, we compare both space-first and feature-first convolution schemes and investigate their effectiveness. Our approach is based on simple linear algebra and analytic solutions, and is highly scalable compared with the current one-vs-one or one-vs-all approach, making it possible to quickly train middle-level features from a number of pairwise part regions. We experimentally show the effectiveness of our method using the standard Caltech-Birds and Stanford-Cars datasets.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2015EDP7358/_p
Copy
@ARTICLE{e99-d_6_1626,
author={Hideki NAKAYAMA, Tomoya TSUDA, },
journal={IEICE TRANSACTIONS on Information},
title={Efficient Two-Step Middle-Level Part Feature Extraction for Fine-Grained Visual Categorization},
year={2016},
volume={E99-D},
number={6},
pages={1626-1634},
abstract={Fine-grained visual categorization (FGVC) has drawn increasing attention as an emerging research field in recent years. In contrast to generic-domain visual recognition, FGVC is characterized by high intra-class and subtle inter-class variations. To distinguish conceptually and visually similar categories, highly discriminative visual features must be extracted. Moreover, FGVC has highly specialized and task-specific nature. It is not always easy to obtain a sufficiently large-scale training dataset. Therefore, the key to success in practical FGVC systems is to efficiently exploit discriminative features from a limited number of training examples. In this paper, we propose an efficient two-step dimensionality compression method to derive compact middle-level part-based features. To do this, we compare both space-first and feature-first convolution schemes and investigate their effectiveness. Our approach is based on simple linear algebra and analytic solutions, and is highly scalable compared with the current one-vs-one or one-vs-all approach, making it possible to quickly train middle-level features from a number of pairwise part regions. We experimentally show the effectiveness of our method using the standard Caltech-Birds and Stanford-Cars datasets.},
keywords={},
doi={10.1587/transinf.2015EDP7358},
ISSN={1745-1361},
month={June},}
Copy
TY - JOUR
TI - Efficient Two-Step Middle-Level Part Feature Extraction for Fine-Grained Visual Categorization
T2 - IEICE TRANSACTIONS on Information
SP - 1626
EP - 1634
AU - Hideki NAKAYAMA
AU - Tomoya TSUDA
PY - 2016
DO - 10.1587/transinf.2015EDP7358
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E99-D
IS - 6
JA - IEICE TRANSACTIONS on Information
Y1 - June 2016
AB - Fine-grained visual categorization (FGVC) has drawn increasing attention as an emerging research field in recent years. In contrast to generic-domain visual recognition, FGVC is characterized by high intra-class and subtle inter-class variations. To distinguish conceptually and visually similar categories, highly discriminative visual features must be extracted. Moreover, FGVC has highly specialized and task-specific nature. It is not always easy to obtain a sufficiently large-scale training dataset. Therefore, the key to success in practical FGVC systems is to efficiently exploit discriminative features from a limited number of training examples. In this paper, we propose an efficient two-step dimensionality compression method to derive compact middle-level part-based features. To do this, we compare both space-first and feature-first convolution schemes and investigate their effectiveness. Our approach is based on simple linear algebra and analytic solutions, and is highly scalable compared with the current one-vs-one or one-vs-all approach, making it possible to quickly train middle-level features from a number of pairwise part regions. We experimentally show the effectiveness of our method using the standard Caltech-Birds and Stanford-Cars datasets.
ER -