Real-Time Polyp Detection: A Speed and Performance Analysis of YOLOv5 and YOLOv6

Öz

Colorectal cancer can potentially be prevented by detecting polyps missed during colonoscopy using a computer aided diagnosis system. Therefore, a diagnostic algorithm, which detects polyps in real-time, was developed to assist endoscopy specialists. You look only once v5 (yolov5) and you look only once v6 (yolov6) models were used for polyp detection. In addition to open-source data, a new private dataset was also used for training object detection models. According to the results, yolov5x and yolov6l achieved mean average precision 50 (mAP50) rates of 0.896 and 0.913, respectively. When yolov5x and yolov6l were compared, it was concluded that yolov5x was better in terms of precision, while yolov6l was better in terms of recall. When models were compared with other studies in the literature, yolov5x outperformed other studies in terms of f1-score with a rate of 0.876 and yolov6l outperformed other studies in terms of recall with a rate of 0.893.

Anahtar Kelimeler

• Real time polyp detection
• Yolov5
• Yolov6
• Object detection
• Computer aided diagnosis

Gerçek Zamanlı Polip Tespiti: YOLOv5 ve YOLOv6'nın Hız ve Performans Analizi

Öz

Kolorektal kanser, kolonoskopi sırasında gözden kaçan poliplerin bilgisayar destekli teşhis sistemi ile tespit edilmesiyle potansiyel olarak önlenebilir. Bu nedenle, endoskopi uzmanlarına yardımcı olmak amacıyla, polipleri gerçek zamanlı olarak tespit eden bir teşhis algoritması geliştirildi. Polip tespiti için you look only once v5 (yolov5) ve you look only once v6 (yolov6) modelleri kullanıldı. Açık kaynaklı verilere ek olarak, nesne tespiti modellerini eğitmek için yeni bir özel veri seti de kullanıldı. Sonuçlara göre, yolov5x ve yolov6l sırasıyla 0.896 ve 0.913 mean average precision 50 (mAP50) oranlarına ulaştı. Yolov5x ve yolov6l karşılaştırıldığında, yolov5x'in hassasiyet açısından daha iyi olduğu, yolov6l'nin ise duyarlılık açısından daha iyi olduğu sonucuna varıldı. Modeller diğer çalışmalardaki sonuçlarla karşılaştırıldığında, yolov5x 0.876 f1-skoru oranıyla diğer çalışmalardan daha iyi performans sergilerken, yolov6l 0.893 duyarlılık oranıyla diğer çalışmaları geride bıraktı.

Anahtar Kelimeler

• Gerçek zamanlı polip tespiti
• Yolov5
• Yolov6
• Bilgisayar destekli tanı
• Nesne tespiti

Kaynakça

1. Bernal J., Sanchez J., Vilarino F. Towards automatic polyp detection with a polyp appearance model. Pattern Recognition 2012; 45(9). 3166-3182.
2. Bernal J., Sanchez FJ., Esparrach GF., Gil D., Rodriguez C., Vilarino F. WM-DOVA maps for accurate polyp highlighting in colonoscopy: Validation vs. saliency maps from physicians. Computerized Medical Imaging and Graphics 2015; 43: 99-111.
3. Bochkovskiy A., Wang CY., Liao HYM. Yolov4: Optimal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934.
4. Buslaev A., Iglovikov V., Khvedchenya E., Parinov A., Druzhinin A., Kalinin M., Alexandr A. Albumentations: fast and flexible image augmentations. Information 2020; 11(2).
5. Chen BL., Wan JJ., Chen TY., Yu YT., Ji M. A self-attention based faster R-CNN for polyp detection from colonoscopy images. Biomedical Signal Processing and Control 2021; 70: 103019.
6. Diederik PK. Adam: A method for stochastic optimization. 2014.
7. Ding X., Zhang X., Ma N., Han J., Ding G., Sun J. RepVGG: making vgg-style convnets great again. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021, 13728-13737, Nashville, TN, USA.
8. Gevorgyan Z. SIoU loss: More powerful learning for bounding box regression. arXiv preprint arXiv:2205.12740.

9. Girshick R., Donahue J., Darrell T., Malik J. Rich feature hierarchies for accurate object detection and semantic segmentation. 2014 IEEE Conference on Computer Vision and Pattern Recognition, 2014, 580-587, Columbus, OH, USA.
10. Girshick R. Fast R-CNN. 2015 IEEE International Conference on Computer Vision (ICCV), 2015, 1440-1448, Santiago, Chile.
11. He K., Zhang S., Ren S., Sun J. Deep residual learning for ımage recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, 770-778, Las Vegas, NV, USA.
12. Jha D., Smedsrud PH., Riegler MA., Halvorsen P., Lange TD., Johansen D., Johansen HD. Kvasir-SEG: A segmented polyp dataset. 2020 Multimedia Modeling, 2020, 451-462, Daejeon, South Korea.
13. Karaman A., Karaboga D., Pacal I., Akay B., Basturk A., Nalbantoglu U., Coskun S., Sahin O. Hyper-parameter optimization of deep learning architectures using artificial bee colony (ABC) algorithm for high performance real-time automatic colorectal cancer (CRC) polyp detection. Applied Intelligence 2022; 53(12): 15603-15620.
14. Kurniawan H., Arief MAA., Manggala B., Lee S., Kim H., Cho BK. Advanced detection of foreign objects in fresh-cut vegetables using YOLOv5. LWT-Food Science and Technology 2024; 212: 116989.
15. Lin TY., Goyal P., Girshick R., He K., Dollar P. Focal loss for dense object detection. 2017 IEEE International Conference on Computer Vision(ICCV), 2017, 2999-3007, Venice, Italy.
16. Liu W., Anguelov D., Erhan D., Szegedy C., Reed S., Fu CY., Berg AC. SSD: Single shot multibox detector. 2016 European Conference on Computer Vision(ECCV), 2016, 21-37, Amsterdam, Netherlands.
17. Li C., Li L., Jiang H., Weng K., Geng Y., Li L., Ke Z., Li Q., Cheng M., Nie W., Li Y., Zhang B., Liang Y., Zhou L., Xu X., Chu X., Wei X., W Xiaolin. Yolov6: A single-stage object detection framework for industrial applications. arXiv preprint arXiv:2209.02976, 55.
18. Liu S., Qi L., Qin H., Shi J., Jia J. Path aggregation network for instance segmentation. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition 2018, 8759-8768, Salt Lake City, UT, USA.
19. Mirza M., Osindero S. Conditional generative adversarial nets. arXiv preprint arXiv:1411.1784.
20. Misra D. Mish: A self regularized non-monotonic activation function. arXiv preprint arXiv:1908.08681.
21. Mushtag F., Ramesh K., Deshmukh S., Ray T., Parimi C., Tandon P., Jha PK. Nuts and bolts: yolo-v5 and image processing based component identification system. Engineering Applications of Artificial Intelligence 2023; 118: 105665.
22. Pacal I., Karaboga D. A robust real-time deep learning based automatic polyp detection system. Computers in Biology and Medicine 2021; 134: 104519.
23. Pacal I., Karaman A., Karaboga D., Akay B., Basturk A., Nalbantoglu U., Coskun S. An efficient real-time colonic polyp detection with yolo algorithms trained by using negative samples and large datasets. Computers in Biology and Medicine 2022; 141: 105031.
24. Qian Z., Jing W., Lv Y., Zhang W. Automatic polyp detection by combining conditional generative adversarial network and modified you-only-look-once. IEEE Sensors Journal 2022; 22(11): 10841-10849.
25. Rahim T., Hassan SA., Shin SY. A deep convolutional neural network for the detection of polyps in colonoscopy images. Biomedical Signal Processing and Control 2021; 68: 102654.
26. Ren S., He K., Girshick R., Sun J. Faster R-CNN: towards real-time object detection with region proposal networks. IEEE Transactions on Pattern Analysis and Machine Intelligence 2017; 39(6): 1137-1149.
27. Redmon J., Divvala S., Girshick R., Farhadi A. You only look once: unified, real-time object detection. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016; 779-788, Las Vegas, NV, USA.
28. Redmon J., Farhadi A. Yolo9000: better, faster, stronger. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, 6517-6525, Honolulu, HI, USA.
29. Redmon J., Farhadi A. Yolov3: an incremental improvement. arXiv preprint arXiv:1804.02767.
30. Rezatofighi H., Tsoi N., Gwak J., Sadeghian A., Reid I., Savarese S. Generalized intersection over union: a metric and a loss for bounding box regression. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2019; 658-666, Long Beach, CA, USA.
31. Silva J., Histace A., Romain O., Dray X., Granado B. Toward embedded detection of polyps in WCE images for early diagnosis of colorectal cancer. International Journal of Computer Assisted Radiology and Surgery 2014; 9: 283-293.
32. Shin Y., Qadir HA., Aabakken L., Bergsland J., Balasingham I. Automatic colon polyp detection using region based deep cnn and post learning approaches. IEEE Access 2018; 6: 40950-40962.
33. Sornapudi S., Meng F., Yi S. Region-based automated localization of colonoscopy and wireless capsule endoscopy polyps. Applied Sciences 2019; 9(12).
34. Sun X., Jia X., Liang Y., Wang M., Chi X. A defect detection method for a boiler inner wall based on an improved yolo-v5 network and data augmentation technologies. IEEE Access 2022; 10: 93845-93853.
35. Tang CP., Hsieh CH., Lin TL. Computer-aided image enhanced endoscopy automated system to boost polyp and adenoma detection accuracy. Diagnostics 2022; 12(4).
36. Taş M., Yılmaz B. Super resolution convolutional neural network based pre-processing for automatic polyp detection in colonoscopy images. Computers and Electrical Engineering 2021; 90: 106959.
37. Wang CY., Liao HYM., Wu YH., Chen PY., Hsieh JW. Cspnet: A new backbone that can enhance learning capability of cnn. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2020, 1571-1580, Seattle, WA, USA.
38. Wang CY., Bochkovskiy A., Liao HYM. Scaled-yolov4: scaling cross stage partial network. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2021; 13024-13033, Nashville, TN, USA.
39. Wang S., Yin Y., Wang D., Lv Z., Wang Y., Jin Y. An interpretable deep neural network for colorectal polyp diagnosis under colonoscopy. Knowledge-Based Systems 2021; 234: 107568.
40. Xiang L., Wang W., Wu L., Chen S., Hu X., Li J., Tang J., Yang J. Generalized focal loss: learning qualified and distributed bounding boxes for dense object detection. Advances in Neural Information Processing Systems 2020; 33: 21002-21012.
41. Xiao Y., Chang A., Wang Y., Huang Y., Yu J., Huo L. Real-time object detection for substation security early-warning with deep neural network based on yolo-V5. 2022 IEEE IAS Global Conference on Emerging Technologies (GlobConET), 2022, 45-50, Arad, Romania.
42. Xu J., Zhao R., Yu Y., Zhang Q., Bian X., Wang J., Ge Z., Qian D. Real-time automatic polyp detection in colonoscopy using feature enhancement module and spatiotemporal similarity correlation unit. Biomedical Signal Processing and Control 2021; 66: 102503.
43. Zhang H., Wang Y., Dayoub F., Sünderhauf N. VarifocalNet: an iou-aware dense object detector. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2021; 8510-8519, Nashville, TN, USA.
44. Zheng Z., Wang P., Liu W., Li J., Ye R. Distance-IoU loss: faster and better learning for bounding box regression. Proceedings of the AAAI Conference on Artificial Intelligence 2020; 34(7): 12993-13000.