Enhancing Object Detection with YOLOv8 Transfer Learning: A VOC2012 Dataset Study

Nan Zhao

2024

Abstract

Object detection has received widespread attention due to its use in a large number of practical scenarios. In this paper, the primary objective is to develop and evaluate a transfer learning-based object detection framework using the you only look once v8 (YOLOv8) model. The study investigates the performance and parameter influence of YOLOv8 when trained on custom datasets through transfer learning methodologies. Firstly, this paper introduces the Vision Object Classes (VOC) 2012 dataset as the primary input data. Subsequently, the YOLOv8 model is configured as the foundational network for feature extraction and object detection tasks. Additionally, transfer learning techniques are applied to enhance the model's generalizability. Thirdly, key parameters are adjusted and compared for thorough analysis. Furthermore, the YOLOv8 predictive performance is assessed and juxtaposed with results obtained from the COCO dataset. The experimental findings highlight the robust performance of YOLOv8 on the VOC2012 dataset. This study offers valuable insights and serves as a reference for researchers in the field, shedding light on effective strategies for object detection using transfer learning approaches.

Paper Citation

in Harvard Style

Zhao N. (2024). Enhancing Object Detection with YOLOv8 Transfer Learning: A VOC2012 Dataset Study. In Proceedings of the 1st International Conference on Engineering Management, Information Technology and Intelligence - Volume 1: EMITI; ISBN 978-989-758-713-9, SciTePress, pages 429-434. DOI: 10.5220/0012939600004508

in Bibtex Style

@conference{emiti24,
author={Nan Zhao},
title={Enhancing Object Detection with YOLOv8 Transfer Learning: A VOC2012 Dataset Study},
booktitle={Proceedings of the 1st International Conference on Engineering Management, Information Technology and Intelligence - Volume 1: EMITI},
year={2024},
pages={429-434},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012939600004508},
isbn={978-989-758-713-9},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Engineering Management, Information Technology and Intelligence - Volume 1: EMITI
TI - Enhancing Object Detection with YOLOv8 Transfer Learning: A VOC2012 Dataset Study
SN - 978-989-758-713-9
AU - Zhao N.
PY - 2024
SP - 429
EP - 434
DO - 10.5220/0012939600004508
PB - SciTePress