Improved Fast YOLO One-Stage Object Detection Algorithm for Detecting Objects in Images
Fidelis Nfwan Gonten1, Ezekwe, Chinwe Genevra2, Otene Patience Unekwuojo3
1Fidelis Nfwan Gonten, Department of Computing Science, Admiralty University of Nigeria/ Abubakar Tafawa Balewa University, (Delta/Bauch), Nigreria.
2Dr. Ezekwe, Chinwe Genevra, Senior Lecturer, Department of Computing Science, Admiralty University of Nigeria, Ibusa (Delta), Nigeria.
3Otene Patience Unekwuojo, Lecturer, Department of Computing Science, Admiralty University of Nigeria, Ibusa (Delta), Nigeria.
Manuscript received on 20 December 2024 | First Revised Manuscript received on 10 April 2025 | Second Revised Manuscript received on 19 April 2025 | Manuscript Accepted on 15 May 2025 | Manuscript published on 30 May 2025 | PP: 1-8 | Volume-12 Issue-5, May 2025 | Retrieval Number: 100.1/ijies.D459714040425 | DOI: 10.35940/ijies.D4597.12050525
Open Access | Editorial and Publishing Policies | Cite | Zenodo | OJS | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Using a Convolutional neural network (CNN) has gained wide recommendation in the research community, especially in vision systems (Object detection). Recently, CNN has achieved various advancements in object detection in images with tremendous accuracy, but it still faces challenges related to high time complexity. A one-stage object detection algorithm, YOLO (You Only Look Once), used for object classification and localisation, performs exceptionally well, especially in real-time object detection. In this study, we propose an improved, fast YOLO CNN-based algorithm for object detection in images. We introduced hard negative mining for resampling and voting to eliminate negative samples, thereby balancing the number of negative and positive samples. A small convolution operation was used in place of the original convolution, which adjusted the parameters and effectively reduced image detection time. The proposed model outperformed Fast YOLO, achieving a precision of 88.32% and a recall of 89.92%, as evaluated on smart city datasets
Keywords: Convolutional Neural Network (CNN), Object Detection, You Only Look Once (YOLO).
Scope of the Article: Computer Vision