Custom Object Detection Using Transfer Learning with Pretrained Models for Improved Detection Techniques

  • Ashwaq Katham Mtasher Assistant Lecturer Ashwaq Katham Mtasher, Department of Community Health, Health and MedicalTechniques College of Kufa, Al_Furat Al_Awsat Technical University (ATU) Iraq
  • Esraa Hassan Jawad Al-wakel Assistant Lecturer Esraa Hassan Jawad Al-wakel, Dep. Of Computer Science, Al-Furat Al-Awsat Technical University (ATU), Kufa, Iraq
Keywords: Custom Object Detection, Transfer Learning, Pre-Trained Models, Deep Learning

Abstract

Custom object detection plays a vital role in computer vision applications. However, developing an accurate and efficient custom object detector requires a substantial amount of labeled training data and significant computational resources. In this research, we propose a custom object detection framework that leverages transfer learning with pre-trained models to improve detection tech-niques.The framework first utilizes a pre-trained deep learning model, such as ResNet or VGGNet, as a feature extractor. The pre-trained model is trained on a large-scale dataset, enabling it to learn high-level features from various objects. By reusing the pre-trained model's convolutional layers, we effectively capture generic features that can be transferred to the custom object detection task.Experimental evaluations on benchmark datasets demonstrate the effectiveness of our ap-proach. The custom object detector achieved superior detection performance compared to tradi-tional methods, especially when the target objects have limited training data. Additionally, our framework significantly reduces the amount of training time and computational resources required, as it leverages pre-trained models as a starting point.

References

Amisse, C., Jijón-Palma, M. E., & Centeno, J. A. S. (2021). Fine-tuning deep learning models for pedestrian detection. Boletim de Ciências Geodésicas, 27.

Amit, Y., Felzenszwalb, P., & Girshick, R. (2021). Object detection. In Computer Vision: A Reference Guide (pp. 875-883). Cham: Springer International Publishing.

Arulprakash, E., & Aruldoss, M. (2022). A study on generic object detection with emphasis on future research directions. Journal of King Saud University-Computer and Information Sciences, 34(9), 7347-7365.

Deng, A., Li, X., Hu, D., Wang, T., Xiong, H., & Xu, C. Z. (2023). Towards Inadequately Pre-trained Models in Transfer Learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 19397-19408).

Ding, Y., Li, H., Chen, K., & Shou, L. (2023, October). TPUF: Enhancing Cross-domain Sequential Recommendation via Transferring Pre-trained User Features. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management (pp. 410-419).

El-Nouby, A., Izacard, G., Touvron, H., Laptev, I., Jegou, H., & Grave, E. (2021). Are large-scale datasets necessary for self-supervised pre-training?. arXiv preprint arXiv:2112.10740.

Liu, X., Yu, W., Liang, F., Griffith, D., & Golmie, N. (2021). Toward deep transfer learning in industrial internet of things. IEEE Internet of Things Journal, 8(15), 12163-12175.

Montalbo, F. J. P. (2020). A Computer-Aided Diagnosis of Brain Tumors Using a Fine-Tuned YOLO-based Model with Transfer Learning. KSII Transactions on Internet & Information Systems, 14(12).

Walambe, R., Marathe, A., & Kotecha, K. (2021). Multiscale object detection from drone imagery using ensemble transfer learning. Drones, 5(3), 66.

Wang, H., Li, H., & Li, B. (2023). VOT: Revolutionizing Speaker Verification with Memory and Attention Mechanisms. arXiv preprint arXiv:2312.16826.

Published
2024-02-02
How to Cite
Mtasher, A. K., & Al-wakel, E. H. J. (2024). Custom Object Detection Using Transfer Learning with Pretrained Models for Improved Detection Techniques . Journal La Multiapp, 5(1), 10-18. https://doi.org/10.37899/journallamultiapp.v5i1.843