Abstract
The biologically inspired model for object recognition, Hierarchical Model and X (HMAX), has attracted considerable attention in recent years. HMAX is robust (i.e., shift- and scale-invariant), but it is sensitive to rotational deformation, which greatly limits its performance in object recognition. The main reason for this is that HMAX lacks an appropriate directional module against rotational deformation, thereby often leading to mismatch. To address this issue, we propose a novel patch-matching method for HMAX called Dominant Orientation Patch Matching (DOPM), which calculates the dominant orientation of the selected patches and implements patch-to-patch matching. In contrast to patch matching with the whole target image (second layer C1) in the conventional HMAX model, which involves huge amounts of redundant information in the feature representation, the DOPM-based HMAX model (D-HMAX) quantizes the C1 layer to patch sets with better distinctiveness, then realizes patch-to-patch matching based on the dominant orientation. To show the effectiveness of D-HMAX, we apply it to object categorization and conduct experiments on the CalTech101, CalTech05, GRAZ01, and GRAZ02 databases. Our experimental results demonstrate that D-HMAX outperforms conventional HMAX and is comparable to existing architectures that have a similar framework.
| Original language | English |
|---|---|
| Journal | Neurocomputing |
| DOIs | |
| Publication status | Accepted/In press - 2015 Jul 7 |
Fingerprint
Keywords
- Classification
- Dominant orientation
- HMAX
- Matching
- Object recognition
- Patch
ASJC Scopus subject areas
- Artificial Intelligence
- Computer Science Applications
- Cognitive Neuroscience
Cite this
Dominant orientation patch matching for HMAX. / Lu, Yan Feng; Zhang, Hua Zhen; Kang, Tae Koo; Lim, Myo Taeg.
In: Neurocomputing, 07.07.2015.Research output: Contribution to journal › Article
}
TY - JOUR
T1 - Dominant orientation patch matching for HMAX
AU - Lu, Yan Feng
AU - Zhang, Hua Zhen
AU - Kang, Tae Koo
AU - Lim, Myo Taeg
PY - 2015/7/7
Y1 - 2015/7/7
N2 - The biologically inspired model for object recognition, Hierarchical Model and X (HMAX), has attracted considerable attention in recent years. HMAX is robust (i.e., shift- and scale-invariant), but it is sensitive to rotational deformation, which greatly limits its performance in object recognition. The main reason for this is that HMAX lacks an appropriate directional module against rotational deformation, thereby often leading to mismatch. To address this issue, we propose a novel patch-matching method for HMAX called Dominant Orientation Patch Matching (DOPM), which calculates the dominant orientation of the selected patches and implements patch-to-patch matching. In contrast to patch matching with the whole target image (second layer C1) in the conventional HMAX model, which involves huge amounts of redundant information in the feature representation, the DOPM-based HMAX model (D-HMAX) quantizes the C1 layer to patch sets with better distinctiveness, then realizes patch-to-patch matching based on the dominant orientation. To show the effectiveness of D-HMAX, we apply it to object categorization and conduct experiments on the CalTech101, CalTech05, GRAZ01, and GRAZ02 databases. Our experimental results demonstrate that D-HMAX outperforms conventional HMAX and is comparable to existing architectures that have a similar framework.
AB - The biologically inspired model for object recognition, Hierarchical Model and X (HMAX), has attracted considerable attention in recent years. HMAX is robust (i.e., shift- and scale-invariant), but it is sensitive to rotational deformation, which greatly limits its performance in object recognition. The main reason for this is that HMAX lacks an appropriate directional module against rotational deformation, thereby often leading to mismatch. To address this issue, we propose a novel patch-matching method for HMAX called Dominant Orientation Patch Matching (DOPM), which calculates the dominant orientation of the selected patches and implements patch-to-patch matching. In contrast to patch matching with the whole target image (second layer C1) in the conventional HMAX model, which involves huge amounts of redundant information in the feature representation, the DOPM-based HMAX model (D-HMAX) quantizes the C1 layer to patch sets with better distinctiveness, then realizes patch-to-patch matching based on the dominant orientation. To show the effectiveness of D-HMAX, we apply it to object categorization and conduct experiments on the CalTech101, CalTech05, GRAZ01, and GRAZ02 databases. Our experimental results demonstrate that D-HMAX outperforms conventional HMAX and is comparable to existing architectures that have a similar framework.
KW - Classification
KW - Dominant orientation
KW - HMAX
KW - Matching
KW - Object recognition
KW - Patch
UR - http://www.scopus.com/inward/record.url?scp=84959887741&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84959887741&partnerID=8YFLogxK
U2 - 10.1016/j.neucom.2016.01.069
DO - 10.1016/j.neucom.2016.01.069
M3 - Article
JO - Neurocomputing
JF - Neurocomputing
SN - 0925-2312
ER -