Crop & Match: RoI Cropping and Feature Matching for Segmentation of Small Objects

Goo Young Moon; Jong Ok Kim

doi:10.1109/ICCE-Asia59966.2023.10326435

Crop & Match: RoI Cropping and Feature Matching for Segmentation of Small Objects

Goo Young Moon^*
, Jong Ok Kim

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

2 Citations (Scopus)

Abstract

In semantic segmentation, predicting RoI (Region of Interest) classes or objects within small regions faces challenges. To address this issue, we propose Crop&Match, a method to combine RoI cropping and feature matching, to enhance model performance of small RoI classes. The cropping ensures the network to focus on small RoI regions by cropping image regions to contain RoI. And feature matching aligns input images with their RoI cropped versions at a representation level, improving predictive consistency across different perspectives on the same image, thereby boosting RoI class recognition.

Original language	English
Title of host publication	2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9798350344318
DOIs	https://doi.org/10.1109/ICCE-Asia59966.2023.10326435
Publication status	Published - 2023
Event	2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023 - Busan, Korea, Republic of Duration: 2023 Oct 23 → 2023 Oct 25

Publication series

Name	2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023

Conference

Conference	2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023
Country/Territory	Korea, Republic of
City	Busan
Period	23/10/23 → 23/10/25

Bibliographical note

Publisher Copyright:
© 2023 IEEE.

Keywords

contrastive learning
disease
plant
semantic segmentation
small object detection

ASJC Scopus subject areas

Computer Networks and Communications
Computer Vision and Pattern Recognition
Electrical and Electronic Engineering
Media Technology
Instrumentation
Artificial Intelligence

Access to Document

10.1109/ICCE-Asia59966.2023.10326435

Cite this

Moon, G. Y., & Kim, J. O. (2023). Crop & Match: RoI Cropping and Feature Matching for Segmentation of Small Objects. In 2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023 (2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICCE-Asia59966.2023.10326435

Crop & Match: RoI Cropping and Feature Matching for Segmentation of Small Objects. / Moon, Goo Young; Kim, Jong Ok.
2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023. Institute of Electrical and Electronics Engineers Inc., 2023. (2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Moon, GY & Kim, JO 2023, Crop & Match: RoI Cropping and Feature Matching for Segmentation of Small Objects. in 2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023. 2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023, Institute of Electrical and Electronics Engineers Inc., 2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023, Busan, Korea, Republic of, 23/10/23. https://doi.org/10.1109/ICCE-Asia59966.2023.10326435

@inproceedings{8677028db14248be81ad476fe4f582e0,

title = "Crop \& Match: RoI Cropping and Feature Matching for Segmentation of Small Objects",

abstract = "In semantic segmentation, predicting RoI (Region of Interest) classes or objects within small regions faces challenges. To address this issue, we propose Crop\&Match, a method to combine RoI cropping and feature matching, to enhance model performance of small RoI classes. The cropping ensures the network to focus on small RoI regions by cropping image regions to contain RoI. And feature matching aligns input images with their RoI cropped versions at a representation level, improving predictive consistency across different perspectives on the same image, thereby boosting RoI class recognition.",

keywords = "contrastive learning, disease, plant, semantic segmentation, small object detection",

author = "Moon, \{Goo Young\} and Kim, \{Jong Ok\}",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023 ; Conference date: 23-10-2023 Through 25-10-2023",

year = "2023",

doi = "10.1109/ICCE-Asia59966.2023.10326435",

language = "English",

series = "2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023",

}

TY - GEN

T1 - Crop & Match

T2 - 2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023

AU - Moon, Goo Young

AU - Kim, Jong Ok

PY - 2023

Y1 - 2023

N2 - In semantic segmentation, predicting RoI (Region of Interest) classes or objects within small regions faces challenges. To address this issue, we propose Crop&Match, a method to combine RoI cropping and feature matching, to enhance model performance of small RoI classes. The cropping ensures the network to focus on small RoI regions by cropping image regions to contain RoI. And feature matching aligns input images with their RoI cropped versions at a representation level, improving predictive consistency across different perspectives on the same image, thereby boosting RoI class recognition.

AB - In semantic segmentation, predicting RoI (Region of Interest) classes or objects within small regions faces challenges. To address this issue, we propose Crop&Match, a method to combine RoI cropping and feature matching, to enhance model performance of small RoI classes. The cropping ensures the network to focus on small RoI regions by cropping image regions to contain RoI. And feature matching aligns input images with their RoI cropped versions at a representation level, improving predictive consistency across different perspectives on the same image, thereby boosting RoI class recognition.

KW - contrastive learning

KW - disease

KW - plant

KW - semantic segmentation

KW - small object detection

UR - https://www.scopus.com/pages/publications/85179891079

U2 - 10.1109/ICCE-Asia59966.2023.10326435

DO - 10.1109/ICCE-Asia59966.2023.10326435

M3 - Conference contribution

AN - SCOPUS:85179891079

T3 - 2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023

BT - 2023 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2023

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 23 October 2023 through 25 October 2023

ER -

Crop & Match: RoI Cropping and Feature Matching for Segmentation of Small Objects

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this