Necaattin BARIŞÇI, Halil Murat ÜNVER, Berker BAYDAN

Hibrit Mask R-CNN Mimarisi Kullanılarak Köpek ve Kedi Tibia Kırık Yerinin Belirlenmesi

Bu çalışmanın amacı Mask R-CNN çatısında kullanılan orjinal omurga yapısını hibrit hale getirerek köpek ve kedi tibia kırıklarındaki kırık bölgelerinin tespitini daha hızlı ve daha yüksek performans ile sağlamaktır. Yapılan hibrit çalışma ile bigisayarlaştırılmış sistem kullanılarak daha yüksek doğruluk oranıyla veteriner hekimlerin tibia üzerindeki kırık teşhislerine yardımcı olması sağlanacaktır. Bu çalışmada üniversitelerden ve kurumlardan elde edilen toplam 518 adet köpek ve kedi kırık tibia kemiği görüntüsü kullanıldı. Bu çalışmanın F1 skor değeri toplam veri seti üzerinde %85.8 olarak bulundu. Çalışmanın köpek veri seti üzerindeki F1 skor değeri %87.8 olarak bulundu. Çalışmanın kedi veri seti üzerindeki F1 skor değeri %77.7 olarak bulundu. Geliştirilen hibrit sistem ile ortalama bir kırık tibia görüntüsündeki kırık yerinin lokalizasyonu 2.88 saniye sürdüğü tespit edildi. Çalışmanın sonuçları, geliştirilen hibrit sistemin orjinal Mask R-CNN mimarisine göre daha başarılı ve hızlı tespitler yaparak hayvan sağlığının korunması açısından faydalı olacağını gösterdi.

Determining the Location of Tibial Fracture of Dog and Cat Using Hybridized Mask R-CNN Architecture

The aim of this study is to hybridize the original backbone structure used in the Mask R-CNN framework, and to detect fracture location indog and cat tibia fractures faster and with higher performance. With the hybrid study, it will be ensured that veterinarians help diagnosefractures on the tibia with higher accuracy by using a computerized system. In this study, a total of 518 dog and cat fracture tibia images thatobtained from universities and institutions were used. F1 score value of this study on total dataset was found to be 85.8%. F1 score valueof this study on dog dataset was found to be 87.8%. F1 score value of this study on cat dataset was found to be 77.7%. With the developedhybrid system, it was determined that the localization of the fracture in an average tibia image took 2.88 seconds. The results of the studyshowed that the hybrid system developed would be beneficial in terms of protecting animal health by making more successful and fasterdetections than the original Mask R-CNN architecture

PDF

___

1. Singh G, Mishra A, Sagar D: An overview of artificial intelligence. SBIT J Sci Technol, 2 (1): 1-4, 2013. DOI: 10.13140/RG.2.2.20660.19840
2. Dhankar M, Walia N: An introduction to artificial intelligence. In, Kumar M, Choudhary R, Pandey SK (Eds): Emerging Trends in Big Data, IoT and Cyber Security. First Impression, 105-108, Maharaja Surajmal Institute. Excellent Publishing Services, New Delhi, 2020.
3. Haenlein M, Kaplan A: A brief history of artificial intelligence: On the past, present, and future of artificial intelligence. Calif Manag Rev, 61 (4): 5-14, 2019. DOI: 10.1177/0008125619864925
4. Yi PH, Hui FK, Ting DSW: Artificial intelligence and radiology: Collaboration is key. J Am Coll Radiol, 15 (5): 781-783, 2018. DOI: 10.1016/j. jacr.2017.12.037
5. Kalmet PHS, Sanduleanu S, Prımakov S, Wu G, Jochems A, Refaee T, Ibrahim A, Hulst LV, Lambin P, Poeze M: Deep learning in fracture detection: A narrative review. Acta Orthop, 91 (2): 215-220, 2020. DOI: 10.1080/17453674.2019.1711323
6. Pak U, Kim C, Ryu U, Sok K, Pak S: A hybrid model based on convolutional neural networks and long short-term memory for ozone concentration prediction. Air Qual Atmos Health, 11 (3): 883-895, 2018. DOI: 10.1007/s11869-018-0585-1
7. Pesapane F, Volonte C, Codari M, Sardanelli F: Artificial intelligence as a medical device in radiology: Ethical and regulatory issues in Europe and the United States. Insights Imaging, 9 (5): 745-753, 2018. DOI: 10.1007/ s13244-018-0645-y
8. Ting DSW, Wu WC, Toth C: Deep learning for retinopathy of prematurity screening. Br J Ophthalmol, 103 (5): 577-579, 2019. DOI: 10.1136/bjophthalmol-2018-313290
9.Cihan P, Gökçe E, Kalıpsız O: A review of machine learning applications in veterinary field. Kafkas Univ Vet Fak Derg, 23 (4): 673-680, 2017. DOI: 10.9775/kvfd.2016.17281
10. Cihan P, Gökçe E, Atakişi O, Kırmızıgül AH, Erdoğan HM: Prediction of immunoglobulin G in lambs with artificial intelligence methods. Kafkas Univ Vet Fak Derg, 27 (1): 21-27, 2021. DOI: 10.9775/kvfd.2020.24642
11. Baydan B, Ünver HM: Detection of tibial fracture in cats and dogs with deep learning. Ankara Univ Vet Fak Derg, 2021 (Article in press). DOI: 10.33988/auvfd.772685
12. Kim DH, MacKinnon T: Artificial intelligence in fracture detection: Transfer learning from deep convolutional neural networks. Clin Radiol, 73 (5): 439-445, 2018. DOI: 10.1016/j.crad.2017.11.015
13. Kvam J, Gangsei LE, Kongsro J, Schistad Solberg AH: The use of deep learning to automate the segmentation of the skeleton from CT volumes of pigs. Transl Anim Sci, 2 (3): 324-335, 2018. DOI: 10.1093/tas/ txy060
14. Baydan B, Ünver HM: Dataset creation and SSD mobilenet V2 performance evaluation for dog tibia fracture detection. In, II. International Ankara Congress of Scientific Research, Ankara, Turkey, 6-8 March, 2020.
15. Krizhevsky A, Sutskever I, Hinton GE: ImageNet classification with deep convolutional neural networks. Commun ACM, 60 (6): 84-90, 2017. DOI: 10.1145/3065386
16. Albawi S, Mohammed TA, Al-Zawi S: Understanding of a convolutional neural network. In, International Conference on Engineering and Technology (ICET). Antalya, Turkey, 21-23 August, 2017, DOI: 10.1109/ icengtechnol.2017.8308186
17. Ting DSW, Peng L, Varadarajan AV, Keane PA, Burlina PM, Chiang MF, Schmettere L, Pasquale LR, Bressler NM, Webster DR, Abramorff M, Wong TY: Deep learning in ophthalmology: The technical and clinical considerations. Prog Retin Eye Res, 72:100759, 2019. DOI: 10.1016/j. preteyeres.2019.04.003
18. He K, Gkioxari G, Dollar P, Girshick R: Mask R-CNN. In, IEEE International Conference on Computer Vision (ICCV). Venice, Italy, 22-29 October 2017. DOI: 10.1109/ICCV.2017.322
19. Gavrishchaka V, Yang Z, Miao R, Senyukova O: Advantages of hybrid deep learning frameworks in applications with limited data. Int J Mach Learn Comput, 8 (6): 549-558, 2018. DOI: 10.18178/ijmlc.2018.8.6.744
20.Villagra A, Alba E, Leguizamon G: A methodology for the hybridization based in active components: The case of cGA and scatter search. Comput Intell Neurosci, 2016:8289237, 2016. DOI: 10.1155/2016/8289237
21. Ding P, Li J, Wang L, Wen M, Guan Y: HYBRID-CNN: An efficient scheme for abnormal flow detection in the SDN-Based Smart Grid. Secur Commun Netw, 2020 (4): 1-20, 2020. DOI: 10.1155/2020/8850550
22. Gülgün OD, Erol H: Medical image classification with hybrid convolutional neural network models. J Comput Sci Technol, 1 (1): 28-41, 2020.
23. Tzutalin: LabelImg. Git code. https://github.com/tzutalin/labelImg; Accessed: 27 February 2020.
24. Chen Q, Gan X, Huang W, Feng J, Shim H: Road damage detection and classification using Mask R-CNN with DenseNet Backbone. Comput Mater Contin, 65 (3): 2201-2215, 2020. DOI: 10.32604/cmc.2020.011191
25. Tychsen-Smith L, Petersson L: Improving object localization with fitness nms and bounded IOU loss. In, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2018. https://arxiv.org/ pdf/1711.00164.pdf; Accessed: 21 July 2020.
26. Battula BP, Balaganesh D: Medical image data classification using deep learning based hybrid model with CNN and Encoder. RIA, 34 (5): 645-652, 2020. DOI: 10.18280/ria.340516
27. Hosny A, Parmar C, Quackenbush J, Scwartz LH, Aerts HJWL: Artificial intelligence in radiology. Nat Rev Cancer, 18 (8): 500-510, 2018. DOI: 10.1038/s41568-018-0016-5
28. Miskovic V: Machıne learning of hybrid classification models for decision support. In, Sinteza 2014. Impact of Internet on Business Activities in Serbia and Worldwide, Belgrade, Singidunum University, Serbia, 318-323, 2014. DOI: 10.15308/SINTEZA-2014-318-323
29. Bemani A, Baghban A, Mosavi A, Shahab S: Estimating CO2-Brine diffusivity using hybrid models of ANFIS and evolutionary algorithms. Eng Appl Comput Fluid Mech, 14 (1): 818-834, 2020. DOI: 10.1080/ 19942060.2020.1774422
30. Jafarifarmand A, Badamchizadeh MA, Khanmohammadi S, Nazari MA, Tazehkand BM: Real-time ocular artifacts removal of EEG data using a hybrid ICA-ANCapproach. Biomed Signal Process Control, 31, 199-210, 2017. DOI: 10.1016/j.bspc.2016.08.006
31.Bharati S, Podder P, Mondal MRH: Hybrid deep learning for detecting lung diseases from X-ray images. Inform Med Unlocked, 20:100391, 2020. DOI: 10.1016/j.imu.2020.100391
32. Golmohammadi M, Torbati AHHN, Diego SL, Obeid I, Picone J: Automatic analysis of EEGs using big data and hybrid deep learning architectures. Front Hum Neurosci, 13:76, 2019. DOI: 10.3389/fnhum. 2019.00076
33. Ma Y, Luo Y: Bone fracture detection through the two-stage system of Crack-Sensitive Convolutional Neural Network, Inform Med Unlocked, 22:100452, 2021. DOI: 10.1016/j.imu.2020.100452
34. Starosolski ZA, Herman Kan J, Annapragada A: CNN-based detection of distal tibial fractures in radiographic images in the setting of open growth plates. In, Medical Imaging 2020: Computer-Aided Diagnosis. Houston, United States, 16-19 February, 2020. DOI: 10.1117/12.2549297
35. Olczak J, Fahlberg N, Maki A, Razavian AS, Jilert A, Stark A, Sköldenberg O, Gordon M: Artificial intelligence for analysing orthopaedic trauma radiographs. Acta Orthop, 88 (6): 581-586, 2017. DOI: 10.1080/17453674.2017.1344459