Yolov9 ve Yolov10 Modellerinin Çeşitli Nesnelerle Karşılaştırmalı Performans Analizi Metodu

Yıl 2025, Cilt: 18 Sayı: 4, 297 - 303, 31.10.2025

Mert Demir

https://doi.org/10.17671/gazibtd.1624632

Öz

Nesne algılama uygulamaları bilgisayarlı görme çalışmalarının temel taşıdır ve bunlardan biri de YOLO'dur. YOLO modeli son dönemde popüler bir hızlı nesne algılama yöntemi olarak çeşitli versiyonlarıyla dikkat çekmektedir. Piyasaya sürülen her YOLO modeli, kullanıcılarda performans ve veri işleme başarısı konusunda merak uyandırmıştır. Bu modelin farklı versiyonları, tek bir nesne üzerinde tespit çalışması yapılarak karşılaştırılır. Ancak tek bir nesneye dayalı bir modeli değerlendirmek, sınırlı bir gözlem çalışmasından başka bir şey değildir. YOLOv10'da nesne algılama puanında iyileşmeler görülüyor. Birçok çalışma, en son model YOLOv10'un YOLOv9' a göre daha fazla nesne tanıma başarısı gösterdiğini söylemektedir, ancak yapılan kapsamlı bir nesne algılama karşılaştırması sonucu YOLOv9'un nesne algılama başarısında en son model YOLOv10'dan daha iyi olduğunu göstermiştir. Nesne tanıma modelleri klasik olarak tek bir nesne resmi üzerinden analiz edilmektedir. Bu da sınırlı bir başarının sınaması olmaktadır. Bu çalışmada klasik yaklaşımlardan farklı olarak her nesne grubuna ait 20 farklı nesneye ait görseller ve performans başarısını gözlemlemek için 50 adet test görüntüsü ile her iki modelin çalışma kriterleri karşılaştırmalı olarak sunulmaktadır. Yapılan 40 deney sonucunda, önerilen ortalama nesne tanıma ölçütü metoduyla YOLOv9 için 72.7 ve YOLOv10 için 64.9’a varan başarı oranı elde edilmiş ve bu modellere ilişkin hata payları %27.3 ve %35.1 olarak hesaplanmıştır.

Anahtar Kelimeler

YOLOv10 , YOLOv9 , performans karşılaştırması , derin öğrenme , nesne algılama

Comparative Performance Analysis Method of Yolov9 and Yolov10 Models with Various Objects

Öz

Comparative Performance Analysis Method of Yolov9 and Yolov10 Models with Various Objects
Araştırma Makalesi/Research Article


Mert DEMİR

Ege Üniversitesi, Fen Bilimleri Enstitüsü, Bilgisayar Mühendisliği, İzmir, Türkiye
[e-mail]
(Geliş/Received:21.01.2025; Kabul/Accepted:19.09.2025)
DOI: 10.17671/gazibtd.1624632


Abstract— Object detection applications are the cornerstone of computer vision studies, and one of them is YOLO. The YOLO model has recently attracted attention with its various versions as a popular fast object detection method. Each YOLO model released has aroused users' curiosity about performance and data processing success. Different versions of this model are compared by performing a detection study on a single object. However, evaluating a model based on a single object is nothing more than a limited observation study. Improvements are seen in object detection score in YOLOv10. Many studies say that the latest model YOLOv10 has better object recognition success than YOLOv9, but a comprehensive object detection comparison has shown that YOLOv9 is better than the latest model YOLOv10 in object detection success. Object recognition models are classically analyzed on a single object image. This is a limited test of success. In this study, unlike classical approaches, images of 20 different objects belonging to each object group and 50 test images are presented to observe the performance success, and the working criteria of both models are presented comparatively. As a result of 40 experiments, with the proposed average object recognition criterion method, success rates of up to 72.7% for YOLOv9 and 64.9% for YOLOv10 were achieved, and the error margins for these models were calculated as 27.3% and 35.1%.

Anahtar Kelimeler

YOLOv10 , YOLOv9 , Performance comparison , Deep learning , Object detection

Kaynakça

• Z. Zou, K. Chen, Z. Shi, Y. Guo, and J. Ye, "Object detection in 20 years: A survey," Proc. IEEE, vol. 111, no. 3, pp. 257-276, 2023. https://doi.org/10.1109/JPROC.2023.3238524.
• Z.-Q. Zhao, P. Zheng, S.-T. Xu, and X. Wu, "Object detection with deep learning: A review," IEEE Trans. Neural Netw. Learn. Syst., vol. 30, no. 11, pp. 3212-3232, 2019. https://doi.org/10.1109/TNNLS.2018.2876865.
• Y. Xiao, Z. Tian, and J. Yu, "A review of object detection based on deep learning," Multimedia Tools and Applications, vol. 79, pp. 23729–23791, 2020. https://doi.org/10.1007/s11042-020-08976-6.
• W. Chen, H. Huang, S. Peng, C. Zhou, and C. Zhang, "YOLO-face: A real-time face detector," Vis. Comput., vol. 37, no. 8, pp. 805-813, 2020. https://doi.org/10.1007/s00371-020-01831-7.
• E. C. Tetila et al., "YOLO performance analysis for real-time detection of soybean pests," Smart Agricultural Technology, vol. 100405, 2024. https://doi.org/10.1016/j.atech.2024.100405.
• G. Oreski, "YOLO*C — Adding context improves YOLO performance," Neurocomputing, vol. 555, p. 126655, 2023. https://doi.org/10.1016/j.neucom.2023.126655.
• P. Jiang, D. Ergu, F. Liu, Y. Cai, and B. Ma, "A review of YOLO algorithm developments," Procedia Comput. Sci., vol. 199, pp. 1066-1073, 2022. https://doi.org/10.1016/j.procs.2022.01.135.
• C. Liu, Y. Tao, J. Liang, K. Li, and Y. Chen, "Object detection based on YOLO network," in 2018 IEEE 4th Inf. Technol. Mechatronics Eng. Conf. (ITOEC), pp. 799-803, 2018. https://doi.org/10.1109/ITOEC.2018.8740604.
• W. Chen, H. Huang, S. Peng et al., "YOLO-face: A real-time face detector," Vis. Comput., vol. 37, pp. 805–813, 2021. https://doi.org/10.1007/s00371-020-01831-7.
• J. Du, "Understanding of object detection based on CNN family and YOLO," J. Phys. Conf. Ser., vol. 1004, 2022. https://doi.org/10.1088/1742-6596/1004/1/012029.
• G. Li, Z. Song, and Q. Fu, "A new method of image detection for small datasets under the framework of YOLO network IEEE 3rd Advanced Information Technology, Electronic and Automation Control Conference (IAEAC)), pp. 1031-1035, 2018. https://doi.org/10.1109/IAEAC.2018.8577214.
• W.-Y. Hsu and W.-Y. Lin, "Ratio-and-scale-aware YOLO for pedestrian detection," IEEE Trans. Image Process., vol. 30, pp. 934-947, 2021. https://doi.org/10.1109/TIP.2020.3039574.
• Z. Yu, H. Huang, W. Chen et al., "YOLO-FaceV2: A scale and occlusion aware face detector," Pattern Recognit., vol. 155, p. 110714, 2022. https://doi.org/10.48550/arXiv.2208.02019.
• S. Xu et al., "PP-YOLOE: An evolved version of YOLO," Computer Science, Computer Vision Pattern Recognition, pp. 1-7, 2022. https://doi.org/10.48550/arXiv.2203.16250.
• E. Chai, L. Ta, Z. Ma, and M. Zhi, "ERF-YOLO: A YOLO algorithm compatible with fewer parameters and higher accuracy," Image and Vision Computing, vol. 116, 2021. https://doi.org/10.1016/j.imavis.2021.104317.
• H. Feng, G. Mu, S. Zhong, P. Zhang, and T. Yuan, "Benchmark analysis of YOLO performance on edge intelligence devices," Cryptography, vol. 6, no. 2, p. 16, 2022. https://doi.org/10.3390/cryptography6020016.
• K. Liu, H. Tang, S. He, Q. Yu, Y. Xiong, and N. Wang, "Performance validation of YOLO variants for object detection," BIC '21: Proceedings of the 2021 International Conference on Bioinformatics and Intelligent Computing, pp. 239–243, 2021. https://doi.org/10.1145/3448748.3448786.
• I. S. Gillani, M. R. Munawar, S. Talha et al., "YOLOv5, YOLO-x, YOLO-r, YOLOv7 performance comparison: A survey," Artificial Intelligence Fuzzy Logic System, pp. 17-28, 2022. http://dx.doi.org/10.5121/csit.2022.121602.
• T. Diwan, G. Anirudh, and J. V. Tembhurne, "Object detection using YOLO: Challenges, architectural successors, datasets and applications," Multimedia Tools Applications, vol. 82, no. 6, pp. 9243-9275, 2023. https://doi.org/10.1007/s11042-022-13644-y.
• D. J. Shin and J. J. Kim, "A deep learning framework performance evaluation to use YOLO in Nvidia Jetson platform," Applied Sciences, vol. 12, no. 8, p. 3734, 2022. https://doi.org/10.3390/app12083734.
• F. Bashir and F. Porikli, "Performance evaluation of object detection and tracking systems," in Proc. 9th IEEE Int. Workshop PETS, pp. 7-14, 2006.
• R. Padilla, S. L. Netto, and E. A. B. da Silva, "A survey on performance metrics for object-detection algorithms," 2020 International Conference on Systems, Signals and Image Processing (IWSSIP), pp. 237-242, 2020. https://doi.org/10.1109/IWSSIP48289.2020.914513.
• M. Bakirci and I. Bayraktar, "Boosting aircraft monitoring and security through ground surveillance optimization with YOLOv9," 12th International Symposium on Digital Forensics and Security (ISDFS), pp. 1-6, San Antonio, TX, USA, 2024. https://doi.org/10.1109/ISDFS60797.2024.1052734.
• R. Sapkota et al., "Comprehensive performance evaluation of YOLOv10, YOLOv9 and YOLOv8 on detecting and counting fruitlet in complex orchard environments," Computer Science, Computer Vision Pattern Recognition, pp. 1-29, 2024. https://doi.org/10.48550/arXiv.2407.12040.
• C. Chien, R. Ju, K. Chou, and J. Chiang, "YOLOv9 for fracture detection in pediatric wrist trauma X-ray images," Electronics Letters, pp. 1-3, 2024. http://dx.doi.org/10.22541/au.171490309.99649889/v1.
• L. Song, Y. Wu, and W. Zhang, "Research on fire detection based on the YOLOv9 algorithm," IEEE 4th International Conference on Electronic Technology, Communication and Information (ICETCI), pp. 1398-1406, Changchun, China, 2024. https://doi.org/10.1109/ICETCI61221.2024.105945.
• V. Breive and T. Sledevic, "Person detection in thermal images: A comparative analysis of YOLOv8 and YOLOv9 models," IEEE Open Conference of Electrical, Electronic and Information Sciences (eStream), pp. 1-4, Vilnius, Lithuania, 2024. https://doi.org/10.1109/eStream61684.2024.1054260.
• A. S. Geetha, M. A. R. Alif, M. Hussain, and P. Allen, "Comparative analysis of YOLOv8 and YOLOv10 in vehicle detection: Performance metrics and model efficacy", Vehicles, 6(3), 1364-1382, 2024. https://doi.org/10.3390/vehicles6030065.
• A. S. Geetha and M. Hussain, "A comparative analysis of YOLOv5, YOLOv8, and YOLOv10 in kitchen safety," arXiv preprint, arXiv:2407.20872, 2024. https://doi.org/10.48550/arXiv.2407.20872.
• M. Hussain and R. Khanam, "In-depth review of YOLOv1 to YOLOv10 variants for enhanced photovoltaic defect detection," Solar, vol. 4, pp. 351-386, 2024. https://doi.org/10.3390/solar4030016.
• A. Vijayakumar and S. Vairavasundaram, "YOLO-based object detection models: A review and its applications," Multimedia Tools and Applicatons, 2024. https://doi.org/10.1007/s11042-024-18872-y.
• Kaggle Datasets- URL: https://www.kaggle.com/datasets (Latest access: 10.01.2025).