KDVGG-Lite: A Distilled Approach for Enhancing the Accuracy of Image Classification

Shahriar Shakir Sumit; Sreenatha Anavatti; Murat Tahtali; Seyedali Mirjalili; Ugur Turhan

doi:10.1007/978-981-97-4985-0_2

KDVGG-Lite: A Distilled Approach for Enhancing the Accuracy of Image Classification

Shahriar Shakir Sumit
, Sreenatha Anavatti
, Murat Tahtali
, Seyedali Mirjalili
, Ugur Turhan

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

Abstract

In recent years, there has been a growing focus on the development of compact and efficient network techniques in the computer vision research field. Towards this goal, this study presents KDVGG-Lite,an innovative image classification model that is inspired from the state of-the-art VGG16 architecture. KDVGG-Lite utilizes knowledge distillation along with pruning technique from the intricate VGG16 model, attaining exceptional results on CIFAR-10 and Fashion-MNIST datasets. Knowledge distillation fusion guarantees efficient compression of essential information, while pruning enhances the model’s design for resource constrained environments. The results demonstrate the exceptional accuracy(an increase of 11.89% and 11.28%), precision (11.51% and 10.06%),recall (11.89% and 11.28%) and F1 score (12.38% and 11.40%) achieved by KDVGG-Lite on CIFAR-10 and Fashion-MNIST respectively, despite having almost 16 times fewer parameters, surpassing its VGG16 counterpart.

Original language	English
Title of host publication	Intelligent Information and Database Systems - 16th Asian Conference, ACIIDS 2024, Proceedings
Editors	Ngoc Thanh Nguyen, Krystian Wojtkiewicz, Richard Chbeir, Yannis Manolopoulos, Hamido Fujita, Tzung-Pei Hong, Le Minh Nguyen
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	15-27
Number of pages	13
ISBN (Print)	9789819749843
DOIs	https://doi.org/10.1007/978-981-97-4985-0_2
Publication status	Published - 2024
Event	16th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2024 - Ras Al Khaimah, United Arab Emirates Duration: 15 Apr 2024 → 18 Apr 2024

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	14796 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	16th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2024
Country/Territory	United Arab Emirates
City	Ras Al Khaimah
Period	15/04/24 → 18/04/24

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 3 Good Health and Well-being
SDG 7 Affordable and Clean Energy
SDG 9 Industry, Innovation, and Infrastructure
SDG 11 Sustainable Cities and Communities
SDG 12 Responsible Consumption and Production
SDG 13 Climate Action
SDG 17 Partnerships for the Goals

Keywords

CIFAR-10
CNNs
Computer Vision
Fashion-MNIST
Image Classification
Knowledge Distillation
Lightweight CNN
VGG

Access to Document

10.1007/978-981-97-4985-0_2

Cite this

Sumit, S. S., Anavatti, S., Tahtali, M., Mirjalili, S., & Turhan, U. (2024). KDVGG-Lite: A Distilled Approach for Enhancing the Accuracy of Image Classification. In N. T. Nguyen, K. Wojtkiewicz, R. Chbeir, Y. Manolopoulos, H. Fujita, T.-P. Hong, & L. M. Nguyen (Eds.), Intelligent Information and Database Systems - 16th Asian Conference, ACIIDS 2024, Proceedings (pp. 15-27). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14796 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-981-97-4985-0_2

Sumit, Shahriar Shakir ; Anavatti, Sreenatha ; Tahtali, Murat et al. / KDVGG-Lite : A Distilled Approach for Enhancing the Accuracy of Image Classification. Intelligent Information and Database Systems - 16th Asian Conference, ACIIDS 2024, Proceedings. editor / Ngoc Thanh Nguyen ; Krystian Wojtkiewicz ; Richard Chbeir ; Yannis Manolopoulos ; Hamido Fujita ; Tzung-Pei Hong ; Le Minh Nguyen. Springer Science and Business Media Deutschland GmbH, 2024. pp. 15-27 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{d6b1b3e4ad2d4860926055b622c721ac,

title = "KDVGG-Lite: A Distilled Approach for Enhancing the Accuracy of Image Classification",

abstract = "In recent years, there has been a growing focus on the development of compact and efficient network techniques in the computer vision research field. Towards this goal, this study presents KDVGG-Lite,an innovative image classification model that is inspired from the state of-the-art VGG16 architecture. KDVGG-Lite utilizes knowledge distillation along with pruning technique from the intricate VGG16 model, attaining exceptional results on CIFAR-10 and Fashion-MNIST datasets. Knowledge distillation fusion guarantees efficient compression of essential information, while pruning enhances the model{\textquoteright}s design for resource constrained environments. The results demonstrate the exceptional accuracy(an increase of 11.89\% and 11.28\%), precision (11.51\% and 10.06\%),recall (11.89\% and 11.28\%) and F1 score (12.38\% and 11.40\%) achieved by KDVGG-Lite on CIFAR-10 and Fashion-MNIST respectively, despite having almost 16 times fewer parameters, surpassing its VGG16 counterpart.",

keywords = "CIFAR-10, CNNs, Computer Vision, Fashion-MNIST, Image Classification, Knowledge Distillation, Lightweight CNN, VGG",

author = "Sumit, \{Shahriar Shakir\} and Sreenatha Anavatti and Murat Tahtali and Seyedali Mirjalili and Ugur Turhan",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.; 16th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2024 ; Conference date: 15-04-2024 Through 18-04-2024",

year = "2024",

doi = "10.1007/978-981-97-4985-0\_2",

language = "English",

isbn = "9789819749843",

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

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "15--27",

editor = "Nguyen, \{Ngoc Thanh\} and Krystian Wojtkiewicz and Richard Chbeir and Yannis Manolopoulos and Hamido Fujita and Tzung-Pei Hong and Nguyen, \{Le Minh\}",

booktitle = "Intelligent Information and Database Systems - 16th Asian Conference, ACIIDS 2024, Proceedings",

}

Sumit, SS, Anavatti, S, Tahtali, M, Mirjalili, S & Turhan, U 2024, KDVGG-Lite: A Distilled Approach for Enhancing the Accuracy of Image Classification. in NT Nguyen, K Wojtkiewicz, R Chbeir, Y Manolopoulos, H Fujita, T-P Hong & LM Nguyen (eds), Intelligent Information and Database Systems - 16th Asian Conference, ACIIDS 2024, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14796 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 15-27, 16th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2024, Ras Al Khaimah, United Arab Emirates, 15/04/24. https://doi.org/10.1007/978-981-97-4985-0_2

KDVGG-Lite: A Distilled Approach for Enhancing the Accuracy of Image Classification. / Sumit, Shahriar Shakir; Anavatti, Sreenatha; Tahtali, Murat et al.
Intelligent Information and Database Systems - 16th Asian Conference, ACIIDS 2024, Proceedings. ed. / Ngoc Thanh Nguyen; Krystian Wojtkiewicz; Richard Chbeir; Yannis Manolopoulos; Hamido Fujita; Tzung-Pei Hong; Le Minh Nguyen. Springer Science and Business Media Deutschland GmbH, 2024. p. 15-27 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14796 LNAI).

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

TY - GEN

T1 - KDVGG-Lite

T2 - 16th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2024

AU - Sumit, Shahriar Shakir

AU - Anavatti, Sreenatha

AU - Tahtali, Murat

AU - Mirjalili, Seyedali

AU - Turhan, Ugur

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

PY - 2024

Y1 - 2024

N2 - In recent years, there has been a growing focus on the development of compact and efficient network techniques in the computer vision research field. Towards this goal, this study presents KDVGG-Lite,an innovative image classification model that is inspired from the state of-the-art VGG16 architecture. KDVGG-Lite utilizes knowledge distillation along with pruning technique from the intricate VGG16 model, attaining exceptional results on CIFAR-10 and Fashion-MNIST datasets. Knowledge distillation fusion guarantees efficient compression of essential information, while pruning enhances the model’s design for resource constrained environments. The results demonstrate the exceptional accuracy(an increase of 11.89% and 11.28%), precision (11.51% and 10.06%),recall (11.89% and 11.28%) and F1 score (12.38% and 11.40%) achieved by KDVGG-Lite on CIFAR-10 and Fashion-MNIST respectively, despite having almost 16 times fewer parameters, surpassing its VGG16 counterpart.

AB - In recent years, there has been a growing focus on the development of compact and efficient network techniques in the computer vision research field. Towards this goal, this study presents KDVGG-Lite,an innovative image classification model that is inspired from the state of-the-art VGG16 architecture. KDVGG-Lite utilizes knowledge distillation along with pruning technique from the intricate VGG16 model, attaining exceptional results on CIFAR-10 and Fashion-MNIST datasets. Knowledge distillation fusion guarantees efficient compression of essential information, while pruning enhances the model’s design for resource constrained environments. The results demonstrate the exceptional accuracy(an increase of 11.89% and 11.28%), precision (11.51% and 10.06%),recall (11.89% and 11.28%) and F1 score (12.38% and 11.40%) achieved by KDVGG-Lite on CIFAR-10 and Fashion-MNIST respectively, despite having almost 16 times fewer parameters, surpassing its VGG16 counterpart.

KW - CIFAR-10

KW - CNNs

KW - Computer Vision

KW - Fashion-MNIST

KW - Image Classification

KW - Knowledge Distillation

KW - Lightweight CNN

KW - VGG

UR - https://www.scopus.com/pages/publications/85210097139

U2 - 10.1007/978-981-97-4985-0_2

DO - 10.1007/978-981-97-4985-0_2

M3 - Conference contribution

AN - SCOPUS:85210097139

SN - 9789819749843

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

SP - 15

EP - 27

BT - Intelligent Information and Database Systems - 16th Asian Conference, ACIIDS 2024, Proceedings

A2 - Nguyen, Ngoc Thanh

A2 - Wojtkiewicz, Krystian

A2 - Chbeir, Richard

A2 - Manolopoulos, Yannis

A2 - Fujita, Hamido

A2 - Hong, Tzung-Pei

A2 - Nguyen, Le Minh

PB - Springer Science and Business Media Deutschland GmbH

Y2 - 15 April 2024 through 18 April 2024

ER -

Sumit SS, Anavatti S, Tahtali M, Mirjalili S, Turhan U. KDVGG-Lite: A Distilled Approach for Enhancing the Accuracy of Image Classification. In Nguyen NT, Wojtkiewicz K, Chbeir R, Manolopoulos Y, Fujita H, Hong TP, Nguyen LM, editors, Intelligent Information and Database Systems - 16th Asian Conference, ACIIDS 2024, Proceedings. Springer Science and Business Media Deutschland GmbH. 2024. p. 15-27. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-981-97-4985-0_2

KDVGG-Lite: A Distilled Approach for Enhancing the Accuracy of Image Classification

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this