Efficient Net-B3 with Convolutional Block Attention and Focal Loss for Road Accident Detection in CCTV Surveillance Videos

• Author(s): M Koteswara Rao; G Raja Sekhar Reddy; P Swarna Kamal; P S C S V Sainadh
• Paper ID: 1717084
• Page: 3725-3731
• Published Date: 01-05-2026
• Published In: Iconic Research And Engineering Journals
• Publisher: IRE Journals
• e-ISSN: 2456-8880
• Volume/Issue: Volume 9 Issue 10 April-2026
• DOI: https://doi.org/10.64388/IREV9I10-1717084

Abstract

Automated road accident detection from CCTV footage is a critical public safety challenge in smart city environments. Prior work using finetuned AlexNet on the ckay16 dataset achieved only 68.0% accuracy, a true positive rate (TPR) of 77.4%, and a critically high false positive rate (FPR) of 42.6% — rendering the system impractical for live deployment. In this paper we present a fully algorithmic framework that addresses these shortcomings without collecting additional data. Our method replaces AlexNet with an EfficientNet-B3 backbone, attaches a Convolutional Block Attention Module (CBAM) to the final feature block, and trains using Focal Loss with label smoothing, a Weighted Random Sampler, and a cosine-annealing learning-rate schedule with warm restarts. At inference, five-crop Test-Time Augmentation (TTA) further reduces false positives. On the held-out ckay16 test set the system achieves 96.0% accuracy, 95.7% TPR, 3.8% FPR, macro-F1 of 96.0%, and ROC-AUC of 0.981 — a 28-point accuracy gain and 39-point FPR reduction over the AlexNet baseline, demonstrating significant improvement over existing baselines.

Keywords

Road accident detection, EfficientNet, CBAM, Attention mechanism, Focal loss, Test-time augmentation, Transfer learning, CCTV surveillance

Citations

IRE Journals:
M Koteswara Rao, G Raja Sekhar Reddy, P Swarna Kamal, P S C S V Sainadh "Efficient Net-B3 with Convolutional Block Attention and Focal Loss for Road Accident Detection in CCTV Surveillance Videos" Iconic Research And Engineering Journals Volume 9 Issue 10 2026 Page 3725-3731 https://doi.org/10.64388/IREV9I10-1717084

IEEE:
M Koteswara Rao, G Raja Sekhar Reddy, P Swarna Kamal, P S C S V Sainadh "Efficient Net-B3 with Convolutional Block Attention and Focal Loss for Road Accident Detection in CCTV Surveillance Videos" Iconic Research And Engineering Journals, vol. 9, no. 10, Apr. 2026, doi: https://doi.org/10.64388/IREV9I10-1717084

APA:
M Koteswara Rao, G Raja Sekhar Reddy, P Swarna Kamal, P S C S V Sainadh (2026). Efficient Net-B3 with Convolutional Block Attention and Focal Loss for Road Accident Detection in CCTV Surveillance Videos. Iconic Research And Engineering Journals, 9(10). doi: https://doi.org/10.64388/IREV9I10-1717084

MLA:
M Koteswara Rao, G Raja Sekhar Reddy, P Swarna Kamal, P S C S V Sainadh "Efficient Net-B3 with Convolutional Block Attention and Focal Loss for Road Accident Detection in CCTV Surveillance Videos" Iconic Research And Engineering Journals, vol. 9, no. 10, Apr. 2026. Crossref, https://doi.org/10.64388/IREV9I10-1717084

BibTeX

@article{1717084,
author = {M Koteswara Rao, G Raja Sekhar Reddy, P Swarna Kamal, P S C S V Sainadh},
title = {Efficient Net-B3 with Convolutional Block Attention and Focal Loss for Road Accident Detection in CCTV Surveillance Videos},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {9},
number = {10},
pages = {3725-3731},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1717084.pdf},
abstract = {Automated road accident detection from CCTV footage is a critical public safety challenge in smart city environments. Prior work using finetuned AlexNet on the ckay16 dataset achieved only 68.0% accuracy, a true positive rate (TPR) of 77.4%, and a critically high false positive rate (FPR) of 42.6% — rendering the system impractical for live deployment. In this paper we present a fully algorithmic framework that addresses these shortcomings without collecting additional data. Our method replaces AlexNet with an EfficientNet-B3 backbone, attaches a Convolutional Block Attention Module (CBAM) to the final feature block, and trains using Focal Loss with label smoothing, a Weighted Random Sampler, and a cosine-annealing learning-rate schedule with warm restarts. At inference, five-crop Test-Time Augmentation (TTA) further reduces false positives. On the held-out ckay16 test set the system achieves 96.0% accuracy, 95.7% TPR, 3.8% FPR, macro-F1 of 96.0%, and ROC-AUC of 0.981 — a 28-point accuracy gain and 39-point FPR reduction over the AlexNet baseline, demonstrating significant improvement over existing baselines.},
keywords = {Road accident detection, EfficientNet, CBAM, Attention mechanism, Focal loss, Test-time augmentation, Transfer learning, CCTV surveillance},
month = {April}
}