Dual-Teacher Color and Structure Distillation for NIR Image Colorization

Tae Sung Park; Jong Ok Kim

doi:10.1109/VCIP59821.2023.10402668

Dual-Teacher Color and Structure Distillation for NIR Image Colorization

Tae Sung Park^*
, Jong Ok Kim

^*Corresponding author for this work

School of Electrical Engineering

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

3 Citations (Scopus)

Abstract

Near-infrared (NIR) imaging can acquire more details and textures with less noise in low-light environments compared to RGB. As a result, it has been widely used in low-light vision scenarios such as CCTV, autonomous driving, and military applications. However, applying NIR images to the human cognitive system and computer vision algorithms is challenging due to the lack of color information. Therefore, it is essential to colorize NIR images into RGB ones. We propose a teacher-student deep network with dual-Teacher knowledge distillation to better estimate the original color and structure information. Specifically, our dual-Teacher network is designed to separately teach distinct knowledge, such as color and structure qualities in an image, to the student. Finally, a color guided structure (CGS) and color embedding (CE) module are applied to fuse color and structure features. Our trained model can retain color consistency and detailed structure information of objects.

Original language	English
Title of host publication	2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9798350359855
DOIs	https://doi.org/10.1109/VCIP59821.2023.10402668
Publication status	Published - 2023
Event	2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023 - Jeju, Korea, Republic of Duration: 2023 Dec 4 → 2023 Dec 7

Publication series

Name	2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023

Conference

Conference	2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023
Country/Territory	Korea, Republic of
City	Jeju
Period	23/12/4 → 23/12/7

Bibliographical note

Publisher Copyright:
© 2023 IEEE.

Keywords

NIR colorization
color distillation
dual-Teacher
multi-band NIR
structure distillation

ASJC Scopus subject areas

Computer Networks and Communications
Computer Vision and Pattern Recognition
Hardware and Architecture
Signal Processing
Media Technology

Access to Document

10.1109/VCIP59821.2023.10402668

Cite this

Park, T. S., & Kim, J. O. (2023). Dual-Teacher Color and Structure Distillation for NIR Image Colorization. In 2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023 (2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/VCIP59821.2023.10402668

Dual-Teacher Color and Structure Distillation for NIR Image Colorization. / Park, Tae Sung; Kim, Jong Ok.
2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023. Institute of Electrical and Electronics Engineers Inc., 2023. (2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023).

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

Park, TS & Kim, JO 2023, Dual-Teacher Color and Structure Distillation for NIR Image Colorization. in 2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023. 2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023, Institute of Electrical and Electronics Engineers Inc., 2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023, Jeju, Korea, Republic of, 23/12/4. https://doi.org/10.1109/VCIP59821.2023.10402668

Park TS, Kim JO. Dual-Teacher Color and Structure Distillation for NIR Image Colorization. In 2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023. Institute of Electrical and Electronics Engineers Inc. 2023. (2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023). doi: 10.1109/VCIP59821.2023.10402668

@inproceedings{041460f79e8745919d3cd40856dfe60e,

title = "Dual-Teacher Color and Structure Distillation for NIR Image Colorization",

abstract = "Near-infrared (NIR) imaging can acquire more details and textures with less noise in low-light environments compared to RGB. As a result, it has been widely used in low-light vision scenarios such as CCTV, autonomous driving, and military applications. However, applying NIR images to the human cognitive system and computer vision algorithms is challenging due to the lack of color information. Therefore, it is essential to colorize NIR images into RGB ones. We propose a teacher-student deep network with dual-Teacher knowledge distillation to better estimate the original color and structure information. Specifically, our dual-Teacher network is designed to separately teach distinct knowledge, such as color and structure qualities in an image, to the student. Finally, a color guided structure (CGS) and color embedding (CE) module are applied to fuse color and structure features. Our trained model can retain color consistency and detailed structure information of objects.",

keywords = "NIR colorization, color distillation, dual-Teacher, multi-band NIR, structure distillation",

author = "Park, \{Tae Sung\} and Kim, \{Jong Ok\}",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023 ; Conference date: 04-12-2023 Through 07-12-2023",

year = "2023",

doi = "10.1109/VCIP59821.2023.10402668",

language = "English",

series = "2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023",

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

booktitle = "2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023",

}

TY - GEN

T1 - Dual-Teacher Color and Structure Distillation for NIR Image Colorization

AU - Park, Tae Sung

AU - Kim, Jong Ok

PY - 2023

Y1 - 2023

N2 - Near-infrared (NIR) imaging can acquire more details and textures with less noise in low-light environments compared to RGB. As a result, it has been widely used in low-light vision scenarios such as CCTV, autonomous driving, and military applications. However, applying NIR images to the human cognitive system and computer vision algorithms is challenging due to the lack of color information. Therefore, it is essential to colorize NIR images into RGB ones. We propose a teacher-student deep network with dual-Teacher knowledge distillation to better estimate the original color and structure information. Specifically, our dual-Teacher network is designed to separately teach distinct knowledge, such as color and structure qualities in an image, to the student. Finally, a color guided structure (CGS) and color embedding (CE) module are applied to fuse color and structure features. Our trained model can retain color consistency and detailed structure information of objects.

AB - Near-infrared (NIR) imaging can acquire more details and textures with less noise in low-light environments compared to RGB. As a result, it has been widely used in low-light vision scenarios such as CCTV, autonomous driving, and military applications. However, applying NIR images to the human cognitive system and computer vision algorithms is challenging due to the lack of color information. Therefore, it is essential to colorize NIR images into RGB ones. We propose a teacher-student deep network with dual-Teacher knowledge distillation to better estimate the original color and structure information. Specifically, our dual-Teacher network is designed to separately teach distinct knowledge, such as color and structure qualities in an image, to the student. Finally, a color guided structure (CGS) and color embedding (CE) module are applied to fuse color and structure features. Our trained model can retain color consistency and detailed structure information of objects.

KW - NIR colorization

KW - color distillation

KW - dual-Teacher

KW - multi-band NIR

KW - structure distillation

UR - https://www.scopus.com/pages/publications/85184852481

U2 - 10.1109/VCIP59821.2023.10402668

DO - 10.1109/VCIP59821.2023.10402668

M3 - Conference contribution

AN - SCOPUS:85184852481

T3 - 2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023

BT - 2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2023 IEEE International Conference on Visual Communications and Image Processing, VCIP 2023

Y2 - 4 December 2023 through 7 December 2023

ER -

Dual-Teacher Color and Structure Distillation for NIR Image Colorization

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this