RecNet: Reinforcement Common Feature Mapping Network for Fast Template Matching in Visible-LWIR Images
We present RecNet, a novel end-to-end CNN approach for fast template matching in cross-spectral images, addressing nonlinear intensity disparities and appearance differences in ground-level imagery through a simple yet effective design. Our key innovation lies in the seamless integration of Zero-mea...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10811895/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841563360581124096 |
|---|---|
| author | Donyung Kim Seungeon Lee Inho Park Geonjong Kim Sungho Kim |
| author_facet | Donyung Kim Seungeon Lee Inho Park Geonjong Kim Sungho Kim |
| author_sort | Donyung Kim |
| collection | DOAJ |
| description | We present RecNet, a novel end-to-end CNN approach for fast template matching in cross-spectral images, addressing nonlinear intensity disparities and appearance differences in ground-level imagery through a simple yet effective design. Our key innovation lies in the seamless integration of Zero-mean normalized cross correlation (ZNCC), which is well-validated for handling nonlinear intensity variations, with CNN-based shape difference learning. Unlike aerial imagery, ground-level cross-spectral matching presents unique challenges due to limited common features, which RecNet effectively addresses. Through comprehensive experiments, including layer-wise visualization analysis and comparative studies with pooling layer combinations, we validate our architecture’s effectiveness. Experiments on KAIST Pedestrian, Log-Gabor Histogram Descriptor(LGHD), and Road Scene datasets demonstrate RecNet’s superior performance and real-time capabilities compared to state-of-the-art methods. Additional evaluations using KAIST nighttime imagery and the M3FD dataset verify RecNet’s generalization capabilities and stability across diverse scenes and conditions, while also identifying limitations and future research directions. |
| format | Article |
| id | doaj-art-4d0a7927f2954ac98f4c3e9078395fbb |
| institution | Kabale University |
| issn | 2169-3536 |
| language | English |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-4d0a7927f2954ac98f4c3e9078395fbb2025-01-03T00:00:51ZengIEEEIEEE Access2169-35362024-01-011219589019590510.1109/ACCESS.2024.352016910811895RecNet: Reinforcement Common Feature Mapping Network for Fast Template Matching in Visible-LWIR ImagesDonyung Kim0https://orcid.org/0009-0009-9693-2358Seungeon Lee1https://orcid.org/0000-0003-0193-6700Inho Park2Geonjong Kim3Sungho Kim4https://orcid.org/0000-0002-5401-2459Department of Electronic Engineering, Yeungnam University, Gyeongsan-si, South KoreaDepartment of Electronic Engineering, Yeungnam University, Gyeongsan-si, South KoreaHanwha Systems, Seoul, South KoreaHanwha Systems, Seoul, South KoreaDepartment of Electronic Engineering, Yeungnam University, Gyeongsan-si, South KoreaWe present RecNet, a novel end-to-end CNN approach for fast template matching in cross-spectral images, addressing nonlinear intensity disparities and appearance differences in ground-level imagery through a simple yet effective design. Our key innovation lies in the seamless integration of Zero-mean normalized cross correlation (ZNCC), which is well-validated for handling nonlinear intensity variations, with CNN-based shape difference learning. Unlike aerial imagery, ground-level cross-spectral matching presents unique challenges due to limited common features, which RecNet effectively addresses. Through comprehensive experiments, including layer-wise visualization analysis and comparative studies with pooling layer combinations, we validate our architecture’s effectiveness. Experiments on KAIST Pedestrian, Log-Gabor Histogram Descriptor(LGHD), and Road Scene datasets demonstrate RecNet’s superior performance and real-time capabilities compared to state-of-the-art methods. Additional evaluations using KAIST nighttime imagery and the M3FD dataset verify RecNet’s generalization capabilities and stability across diverse scenes and conditions, while also identifying limitations and future research directions.https://ieeexplore.ieee.org/document/10811895/Visible-LWIR template matchingmulti-modalitycross-spectral image processing |
| spellingShingle | Donyung Kim Seungeon Lee Inho Park Geonjong Kim Sungho Kim RecNet: Reinforcement Common Feature Mapping Network for Fast Template Matching in Visible-LWIR Images IEEE Access Visible-LWIR template matching multi-modality cross-spectral image processing |
| title | RecNet: Reinforcement Common Feature Mapping Network for Fast Template Matching in Visible-LWIR Images |
| title_full | RecNet: Reinforcement Common Feature Mapping Network for Fast Template Matching in Visible-LWIR Images |
| title_fullStr | RecNet: Reinforcement Common Feature Mapping Network for Fast Template Matching in Visible-LWIR Images |
| title_full_unstemmed | RecNet: Reinforcement Common Feature Mapping Network for Fast Template Matching in Visible-LWIR Images |
| title_short | RecNet: Reinforcement Common Feature Mapping Network for Fast Template Matching in Visible-LWIR Images |
| title_sort | recnet reinforcement common feature mapping network for fast template matching in visible lwir images |
| topic | Visible-LWIR template matching multi-modality cross-spectral image processing |
| url | https://ieeexplore.ieee.org/document/10811895/ |
| work_keys_str_mv | AT donyungkim recnetreinforcementcommonfeaturemappingnetworkforfasttemplatematchinginvisiblelwirimages AT seungeonlee recnetreinforcementcommonfeaturemappingnetworkforfasttemplatematchinginvisiblelwirimages AT inhopark recnetreinforcementcommonfeaturemappingnetworkforfasttemplatematchinginvisiblelwirimages AT geonjongkim recnetreinforcementcommonfeaturemappingnetworkforfasttemplatematchinginvisiblelwirimages AT sunghokim recnetreinforcementcommonfeaturemappingnetworkforfasttemplatematchinginvisiblelwirimages |