CPE COIN++: Towards Optimized Implicit Neural Representation Compression Via Chebyshev Positional Encoding

Haocheng Chu; Shaohui Dai; Wenqi Ding; Xin Shi; Tianshuo Xu; Pingyang Dai; Shengchuan Zhang; Yan Zhang; Xiang Chang; Chih Min Lin; Fei Chao; Changjiang Shang; Qiang Shen

doi:10.1007/978-981-97-8692-3_36

CPE COIN++: Towards Optimized Implicit Neural Representation Compression Via Chebyshev Positional Encoding

Haocheng Chu, Shaohui Dai, Wenqi Ding, Xin Shi, Tianshuo Xu, Pingyang Dai, Shengchuan Zhang, Yan Zhang, Xiang Chang, Chih Min Lin, Fei Chao^*, Changjiang Shang, Qiang Shen

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference Proceeding (Non-Journal item)

Abstract

COIN++ is a special variant of Implicit Neural Representation (INR), which encodes signals as modulations applied to the base INR network. It is becoming a promising method for applications in image compression. However, INR’s effectiveness is hindered by its inability to capture high-frequency details in the image representation. We propose a novel COIN++ framework using Chebyshev approximation to enhance high-frequency signal learning and image compression. In addition, we design an adaptable image partitioning technology and an integrated quantization method to further the image compression performance of COIN++ in the framework. Experiments demonstrate our framework significantly enhances both representational capacity and compression rate compared to the COIN++ baseline, with notable PSNR improvements.

Original language	English
Title of host publication	Pattern Recognition and Computer Vision - 7th Chinese Conference, PRCV 2024, Proceedings
Editors	Zhouchen Lin, Hongbin Zha, Ming-Ming Cheng, Ran He, Cheng-Lin Liu, Kurban Ubul, Wushouer Silamu, Jie Zhou
Publisher	Springer Nature
Pages	509-524
Number of pages	16
ISBN (Print)	9789819786916
DOIs	https://doi.org/10.1007/978-981-97-8692-3_36
Publication status	Accepted/In press - 25 Jun 2024
Event	7th Chinese Conference on Pattern Recognition and Computer Vision, PRCV 2024 - Urumqi, China Duration: 18 Oct 2024 → 20 Oct 2024

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	15039 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	7th Chinese Conference on Pattern Recognition and Computer Vision, PRCV 2024
Country/Territory	China
City	Urumqi
Period	18 Oct 2024 → 20 Oct 2024

Keywords

Chebyshev approximation
COIN++
Implicit Neural Representation

Access to Document

10.1007/978-981-97-8692-3_36

Permanent link

Persistent link

Cite this

Chu, H., Dai, S., Ding, W., Shi, X., Xu, T., Dai, P., Zhang, S., Zhang, Y., Chang, X., Lin, C. M., Chao, F., Shang, C., & Shen, Q. (Accepted/In press). CPE COIN++: Towards Optimized Implicit Neural Representation Compression Via Chebyshev Positional Encoding. In Z. Lin, H. Zha, M.-M. Cheng, R. He, C.-L. Liu, K. Ubul, W. Silamu, & J. Zhou (Eds.), Pattern Recognition and Computer Vision - 7th Chinese Conference, PRCV 2024, Proceedings (pp. 509-524). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 15039 LNCS). Springer Nature. https://doi.org/10.1007/978-981-97-8692-3_36

Chu, Haocheng ; Dai, Shaohui ; Ding, Wenqi et al. / CPE COIN++ : Towards Optimized Implicit Neural Representation Compression Via Chebyshev Positional Encoding. Pattern Recognition and Computer Vision - 7th Chinese Conference, PRCV 2024, Proceedings. editor / Zhouchen Lin ; Hongbin Zha ; Ming-Ming Cheng ; Ran He ; Cheng-Lin Liu ; Kurban Ubul ; Wushouer Silamu ; Jie Zhou. Springer Nature, 2025. pp. 509-524 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{9da9cfe251f64c52ac0c88878c778705,

title = "CPE COIN++: Towards Optimized Implicit Neural Representation Compression Via Chebyshev Positional Encoding",

abstract = "COIN++ is a special variant of Implicit Neural Representation (INR), which encodes signals as modulations applied to the base INR network. It is becoming a promising method for applications in image compression. However, INR{\textquoteright}s effectiveness is hindered by its inability to capture high-frequency details in the image representation. We propose a novel COIN++ framework using Chebyshev approximation to enhance high-frequency signal learning and image compression. In addition, we design an adaptable image partitioning technology and an integrated quantization method to further the image compression performance of COIN++ in the framework. Experiments demonstrate our framework significantly enhances both representational capacity and compression rate compared to the COIN++ baseline, with notable PSNR improvements.",

keywords = "Chebyshev approximation, COIN++, Implicit Neural Representation",

author = "Haocheng Chu and Shaohui Dai and Wenqi Ding and Xin Shi and Tianshuo Xu and Pingyang Dai and Shengchuan Zhang and Yan Zhang and Xiang Chang and Lin, {Chih Min} and Fei Chao and Changjiang Shang and Qiang Shen",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.; 7th Chinese Conference on Pattern Recognition and Computer Vision, PRCV 2024 ; Conference date: 18-10-2024 Through 20-10-2024",

year = "2024",

month = jun,

day = "25",

doi = "10.1007/978-981-97-8692-3_36",

language = "English",

isbn = "9789819786916",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Nature",

pages = "509--524",

editor = "Zhouchen Lin and Hongbin Zha and Ming-Ming Cheng and Ran He and Cheng-Lin Liu and Kurban Ubul and Wushouer Silamu and Jie Zhou",

booktitle = "Pattern Recognition and Computer Vision - 7th Chinese Conference, PRCV 2024, Proceedings",

address = "Switzerland",

}

Chu, H, Dai, S, Ding, W, Shi, X, Xu, T, Dai, P, Zhang, S, Zhang, Y, Chang, X, Lin, CM, Chao, F , Shang, C & Shen, Q 2025, CPE COIN++: Towards Optimized Implicit Neural Representation Compression Via Chebyshev Positional Encoding. in Z Lin, H Zha, M-M Cheng, R He, C-L Liu, K Ubul, W Silamu & J Zhou (eds), Pattern Recognition and Computer Vision - 7th Chinese Conference, PRCV 2024, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 15039 LNCS, Springer Nature, pp. 509-524, 7th Chinese Conference on Pattern Recognition and Computer Vision, PRCV 2024, Urumqi, China, 18 Oct 2024. https://doi.org/10.1007/978-981-97-8692-3_36

CPE COIN++: Towards Optimized Implicit Neural Representation Compression Via Chebyshev Positional Encoding. / Chu, Haocheng; Dai, Shaohui; Ding, Wenqi et al.
Pattern Recognition and Computer Vision - 7th Chinese Conference, PRCV 2024, Proceedings. ed. / Zhouchen Lin; Hongbin Zha; Ming-Ming Cheng; Ran He; Cheng-Lin Liu; Kurban Ubul; Wushouer Silamu; Jie Zhou. Springer Nature, 2025. p. 509-524 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 15039 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference Proceeding (Non-Journal item)

TY - GEN

T1 - CPE COIN++

T2 - 7th Chinese Conference on Pattern Recognition and Computer Vision, PRCV 2024

AU - Chu, Haocheng

AU - Dai, Shaohui

AU - Ding, Wenqi

AU - Shi, Xin

AU - Xu, Tianshuo

AU - Dai, Pingyang

AU - Zhang, Shengchuan

AU - Zhang, Yan

AU - Chang, Xiang

AU - Lin, Chih Min

AU - Chao, Fei

AU - Shang, Changjiang

AU - Shen, Qiang

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

PY - 2024/6/25

Y1 - 2024/6/25

N2 - COIN++ is a special variant of Implicit Neural Representation (INR), which encodes signals as modulations applied to the base INR network. It is becoming a promising method for applications in image compression. However, INR’s effectiveness is hindered by its inability to capture high-frequency details in the image representation. We propose a novel COIN++ framework using Chebyshev approximation to enhance high-frequency signal learning and image compression. In addition, we design an adaptable image partitioning technology and an integrated quantization method to further the image compression performance of COIN++ in the framework. Experiments demonstrate our framework significantly enhances both representational capacity and compression rate compared to the COIN++ baseline, with notable PSNR improvements.

AB - COIN++ is a special variant of Implicit Neural Representation (INR), which encodes signals as modulations applied to the base INR network. It is becoming a promising method for applications in image compression. However, INR’s effectiveness is hindered by its inability to capture high-frequency details in the image representation. We propose a novel COIN++ framework using Chebyshev approximation to enhance high-frequency signal learning and image compression. In addition, we design an adaptable image partitioning technology and an integrated quantization method to further the image compression performance of COIN++ in the framework. Experiments demonstrate our framework significantly enhances both representational capacity and compression rate compared to the COIN++ baseline, with notable PSNR improvements.

KW - Chebyshev approximation

KW - COIN++

KW - Implicit Neural Representation

UR - http://www.scopus.com/inward/record.url?scp=85209794204&partnerID=8YFLogxK

U2 - 10.1007/978-981-97-8692-3_36

DO - 10.1007/978-981-97-8692-3_36

M3 - Conference Proceeding (Non-Journal item)

AN - SCOPUS:85209794204

SN - 9789819786916

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 509

EP - 524

BT - Pattern Recognition and Computer Vision - 7th Chinese Conference, PRCV 2024, Proceedings

A2 - Lin, Zhouchen

A2 - Zha, Hongbin

A2 - Cheng, Ming-Ming

A2 - He, Ran

A2 - Liu, Cheng-Lin

A2 - Ubul, Kurban

A2 - Silamu, Wushouer

A2 - Zhou, Jie

PB - Springer Nature

Y2 - 18 October 2024 through 20 October 2024

ER -

Chu H, Dai S, Ding W, Shi X, Xu T, Dai P et al. CPE COIN++: Towards Optimized Implicit Neural Representation Compression Via Chebyshev Positional Encoding. In Lin Z, Zha H, Cheng MM, He R, Liu CL, Ubul K, Silamu W, Zhou J, editors, Pattern Recognition and Computer Vision - 7th Chinese Conference, PRCV 2024, Proceedings. Springer Nature. 2025. p. 509-524. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-981-97-8692-3_36