Assessing the efficacy of VGG16 with CNN in non-clinical ASD prediction

Ranjeet Vasant Bidwe; Rashmi Ashtagi; Sashikala Mishra; Simi Bajaj

doi:10.1007/978-981-95-0375-9_13

Assessing the efficacy of VGG16 with CNN in non-clinical ASD prediction

Ranjeet Vasant Bidwe
, Rashmi Ashtagi
, Sashikala Mishra
, Simi Bajaj

Symbiosis International University
MIT World Peace University

Research output: Chapter in Book / Conference Paper › Chapter › peer-review

Abstract

In this study, we introduce a new technique based on image processing to predict “autism spectrum disorder” (ASD) images using a combination of transfer learning and VGG16 Convolutional Neural Network (CNN) architecture. Using an extensive dataset of 2540 training pics, 100 validation, and 300 test images, the model showed an accuracy of 71.33%, which is quite impressive. The proposed approach, by leveraging pre-trained weights on a large image dataset and fine-tuning ASD-specific data, holds great promise for effectively screening people with ASD. Conclusions The results indicate that deep learning methods could help to enhance the accuracy of ASD diagnosis and interventions, inherently improve early diagnosis and specialized therapy interventions for patients with ASD. In general, this research sets the stage for better and more individualized treatments for people with ASD. It also sets the ground for future research focusing on increasing the accuracy and robustness of machine learning models that could be used in ASD prediction and management.

Original language	English
Title of host publication	Information System Design: AI and ML Applications: Proceedings of Ninth International Conference on Information System Design and Intelligent Applications (ISDIA 2025), Volume 4
Editors	Vikrant Bhateja, Soly Mathew Biju, Siba K. Udgata
Place of Publication	Singapore
Publisher	Springer
Pages	155-168
Number of pages	14
ISBN (Electronic)	9789819503759
ISBN (Print)	9789819503742
DOIs	https://doi.org/10.1007/978-981-95-0375-9_13
Publication status	Published - 2026
Event	International Conference on Information System Design and Intelligent Applications - Dubai, United Arab Emirates Duration: 3 Jan 2025 → 4 Jan 2025 Conference number: 9th

Publication series

Name	Lecture Notes in Networks and Systems
Volume	1565 LNNS
ISSN (Print)	2367-3370
ISSN (Electronic)	2367-3389

Conference

Conference	International Conference on Information System Design and Intelligent Applications
Abbreviated title	ISDIA
Country/Territory	United Arab Emirates
City	Dubai
Period	3/01/25 → 4/01/25

Keywords

Convolutional neural network
Deep learning
Eye positioning
Image classification
Transfer learning
VGG16

Access to Document

10.1007/978-981-95-0375-9_13

Cite this

Bidwe, R. V., Ashtagi, R., Mishra, S., & Bajaj, S. (2026). Assessing the efficacy of VGG16 with CNN in non-clinical ASD prediction. In V. Bhateja, S. M. Biju, & S. K. Udgata (Eds.), Information System Design: AI and ML Applications: Proceedings of Ninth International Conference on Information System Design and Intelligent Applications (ISDIA 2025), Volume 4 (pp. 155-168). (Lecture Notes in Networks and Systems; Vol. 1565 LNNS). Springer. https://doi.org/10.1007/978-981-95-0375-9_13

Bidwe, Ranjeet Vasant ; Ashtagi, Rashmi ; Mishra, Sashikala et al. / Assessing the efficacy of VGG16 with CNN in non-clinical ASD prediction. Information System Design: AI and ML Applications: Proceedings of Ninth International Conference on Information System Design and Intelligent Applications (ISDIA 2025), Volume 4. editor / Vikrant Bhateja ; Soly Mathew Biju ; Siba K. Udgata. Singapore : Springer, 2026. pp. 155-168 (Lecture Notes in Networks and Systems).

@inbook{91cac008c42c45c8b19170a024532948,

title = "Assessing the efficacy of VGG16 with CNN in non-clinical ASD prediction",

abstract = "In this study, we introduce a new technique based on image processing to predict “autism spectrum disorder” (ASD) images using a combination of transfer learning and VGG16 Convolutional Neural Network (CNN) architecture. Using an extensive dataset of 2540 training pics, 100 validation, and 300 test images, the model showed an accuracy of 71.33\%, which is quite impressive. The proposed approach, by leveraging pre-trained weights on a large image dataset and fine-tuning ASD-specific data, holds great promise for effectively screening people with ASD. Conclusions The results indicate that deep learning methods could help to enhance the accuracy of ASD diagnosis and interventions, inherently improve early diagnosis and specialized therapy interventions for patients with ASD. In general, this research sets the stage for better and more individualized treatments for people with ASD. It also sets the ground for future research focusing on increasing the accuracy and robustness of machine learning models that could be used in ASD prediction and management.",

keywords = "Convolutional neural network, Deep learning, Eye positioning, Image classification, Transfer learning, VGG16",

author = "Bidwe, \{Ranjeet Vasant\} and Rashmi Ashtagi and Sashikala Mishra and Simi Bajaj",

year = "2026",

doi = "10.1007/978-981-95-0375-9\_13",

language = "English",

isbn = "9789819503742",

series = "Lecture Notes in Networks and Systems",

publisher = "Springer",

pages = "155--168",

editor = "Vikrant Bhateja and Biju, \{Soly Mathew\} and Udgata, \{Siba K.\}",

booktitle = "Information System Design: AI and ML Applications: Proceedings of Ninth International Conference on Information System Design and Intelligent Applications (ISDIA 2025), Volume 4",

note = "International Conference on Information System Design and Intelligent Applications, ISDIA ; Conference date: 03-01-2025 Through 04-01-2025",

}

Bidwe, RV, Ashtagi, R, Mishra, S & Bajaj, S 2026, Assessing the efficacy of VGG16 with CNN in non-clinical ASD prediction. in V Bhateja, SM Biju & SK Udgata (eds), Information System Design: AI and ML Applications: Proceedings of Ninth International Conference on Information System Design and Intelligent Applications (ISDIA 2025), Volume 4. Lecture Notes in Networks and Systems, vol. 1565 LNNS, Springer, Singapore, pp. 155-168, International Conference on Information System Design and Intelligent Applications, Dubai, United Arab Emirates, 3/01/25. https://doi.org/10.1007/978-981-95-0375-9_13

Assessing the efficacy of VGG16 with CNN in non-clinical ASD prediction. / Bidwe, Ranjeet Vasant; Ashtagi, Rashmi; Mishra, Sashikala et al.
Information System Design: AI and ML Applications: Proceedings of Ninth International Conference on Information System Design and Intelligent Applications (ISDIA 2025), Volume 4. ed. / Vikrant Bhateja; Soly Mathew Biju; Siba K. Udgata. Singapore: Springer, 2026. p. 155-168 (Lecture Notes in Networks and Systems; Vol. 1565 LNNS).

Research output: Chapter in Book / Conference Paper › Chapter › peer-review

TY - CHAP

T1 - Assessing the efficacy of VGG16 with CNN in non-clinical ASD prediction

AU - Bidwe, Ranjeet Vasant

AU - Ashtagi, Rashmi

AU - Mishra, Sashikala

AU - Bajaj, Simi

N1 - Conference code: 9th

PY - 2026

Y1 - 2026

N2 - In this study, we introduce a new technique based on image processing to predict “autism spectrum disorder” (ASD) images using a combination of transfer learning and VGG16 Convolutional Neural Network (CNN) architecture. Using an extensive dataset of 2540 training pics, 100 validation, and 300 test images, the model showed an accuracy of 71.33%, which is quite impressive. The proposed approach, by leveraging pre-trained weights on a large image dataset and fine-tuning ASD-specific data, holds great promise for effectively screening people with ASD. Conclusions The results indicate that deep learning methods could help to enhance the accuracy of ASD diagnosis and interventions, inherently improve early diagnosis and specialized therapy interventions for patients with ASD. In general, this research sets the stage for better and more individualized treatments for people with ASD. It also sets the ground for future research focusing on increasing the accuracy and robustness of machine learning models that could be used in ASD prediction and management.

AB - In this study, we introduce a new technique based on image processing to predict “autism spectrum disorder” (ASD) images using a combination of transfer learning and VGG16 Convolutional Neural Network (CNN) architecture. Using an extensive dataset of 2540 training pics, 100 validation, and 300 test images, the model showed an accuracy of 71.33%, which is quite impressive. The proposed approach, by leveraging pre-trained weights on a large image dataset and fine-tuning ASD-specific data, holds great promise for effectively screening people with ASD. Conclusions The results indicate that deep learning methods could help to enhance the accuracy of ASD diagnosis and interventions, inherently improve early diagnosis and specialized therapy interventions for patients with ASD. In general, this research sets the stage for better and more individualized treatments for people with ASD. It also sets the ground for future research focusing on increasing the accuracy and robustness of machine learning models that could be used in ASD prediction and management.

KW - Convolutional neural network

KW - Deep learning

KW - Eye positioning

KW - Image classification

KW - Transfer learning

KW - VGG16

UR - https://www.scopus.com/pages/publications/105027933007

UR - https://go.openathens.net/redirector/westernsydney.edu.au?url=https://doi.org/10.1007/978-981-95-0375-9_13

U2 - 10.1007/978-981-95-0375-9_13

DO - 10.1007/978-981-95-0375-9_13

M3 - Chapter

AN - SCOPUS:105027933007

SN - 9789819503742

T3 - Lecture Notes in Networks and Systems

SP - 155

EP - 168

BT - Information System Design: AI and ML Applications: Proceedings of Ninth International Conference on Information System Design and Intelligent Applications (ISDIA 2025), Volume 4

A2 - Bhateja, Vikrant

A2 - Biju, Soly Mathew

A2 - Udgata, Siba K.

PB - Springer

CY - Singapore

T2 - International Conference on Information System Design and Intelligent Applications

Y2 - 3 January 2025 through 4 January 2025

ER -

Bidwe RV, Ashtagi R, Mishra S, Bajaj S. Assessing the efficacy of VGG16 with CNN in non-clinical ASD prediction. In Bhateja V, Biju SM, Udgata SK, editors, Information System Design: AI and ML Applications: Proceedings of Ninth International Conference on Information System Design and Intelligent Applications (ISDIA 2025), Volume 4. Singapore: Springer. 2026. p. 155-168. (Lecture Notes in Networks and Systems). doi: 10.1007/978-981-95-0375-9_13

Assessing the efficacy of VGG16 with CNN in non-clinical ASD prediction

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this