APPLICATION OF VGGNET AND CBAM ATTENTION MECHANISM TO ENDOSCOPIC BLADDER TISSUE IMAGES

Öz

Convolutional neural networks (CNNs) are widely used to highlight relevant information and suppress irrelevant information in tasks to be performed for detecting, diagnosing, and monitoring medical diseases. However, reasons such as computational complexity, feature quality problems, and increasing feature size challenge the disease detection performance of CNN models. In recent times, attention mechanisms have been used to enhance the performance of CNN models, helping to overcome challenges. The Convolutional Block Attention Module (CBAM) attention mechanism is a method developed to extract relevant complex features with the help of the modules it contains. The integration of this mechanism with the CNN model significantly improves the performance of the model. This study aims to classify bladder tissue with models created by combining VGGNet and CBAM attention mechanism. Experiments were carried out using accuracy, precision, recall, F1-score, and area under the curve metrics (AUC) to compare the performances of VGGNet and VGGNet+CBAM models. According to the results, the VGG19+CBAM model has demonstrated the highest performance values in terms of accuracy, precision, recall, F1-score, and AUC criteria. The accuracy, precision, recall, F1-score, and AUC values of this model are 0.990, 0.992, 0.984, 0.986, and 0.994, respectively. VGGNet+CBAM models have shown better performance than VGGNet models. Based on the achieved performance values, it is evident that the proposed approach is effective in the diagnosis of bladder tumors.

Anahtar Kelimeler

• Convolutional neural network
• CBAM attention mechanism
• endoscopic bladder tissue

VGGNET VE CBAM DİKKAT MEKANİZMASININ ENDOSKOPİK MESANE DOKU GÖRÜNTÜLERİNE UYGULANMASI

Öz

Tıbbi hastalıkların tespiti, tanısı ve izlenmesi amacıyla gerçekleştirilecek görevlerde, ilgili bilgilerin öne çıkartılması ve ilgisiz bilgilerin bastırılmasında Evrişimsel sinir ağları (ESA) yaygın olarak kullanılmaktadır. Ancak ESA modellerinin hesaplama karmaşıklığı, özellik kalitesi sorunu ve artan özellik boyutu gibi nedenler hastalığın tespit performansını zorlamaktadır. Son zamanlarda, ESA modellerinin performansını artırmak için dikkat mekanizmaları kullanılmaktadır, bu da sorunların üstesinden gelmeye yardımcı olmaktadır. Evrişimsel Blok Dikkat Modülü (CBAM) dikkat mekanizması, içerisinde barındırdığı modüller yardımıyla ilgili karmaşık özellikleri çıkarmak için geliştirilmiş bir yöntemdir. ESA modeliyle bu mekanizmanın birleştirilmesi, modelin performansını önemli ölçüde iyileştirilir. Çalışmanın amacı VGGNet ve CBAM dikkat mekanizmasını birleştirerek oluşturulan modellerle mesane tümörünün sınıflandırılmasıdır. Çalışmada VGGNet ve VGGNet+CBAM modellerinin performanslarını karşılaştırmak için doğruluk, kesinlik, duyarlılık, F1-skor ve eğri altında kalan alan (AUC) metrikleri kullanılarak deneyler gerçekleştirilmiştir. Sonuçlara göre VGG19+CBAM modeli, doğruluk, kesinlik, duyarlılık, F1-skor ve AUC ölçütleri açısından en yüksek performans değerlerini göstermiştir. Bu modelin doğruluk, kesinlik, duyarlılık, F1-skor ve AUC ölçütlerinin değerleri sırasıyla 0,990, 0,992, 0,984, 0,986 ve 0,994’tür. VGGNet+CBAM modelleri, VGGNet modellerinden daha iyi performans göstermiştir. Elde edilen performans değerlerine dayanarak, önerilen yaklaşımın mesane tümörü teşhisinde etkili olduğu görülmektedir.

Anahtar Kelimeler

• Evrişimsel sinir ağı
• CBAM dikkat mekanizması
• endoskopik mesane dokusu

Kaynakça

1. Alirezazadeh, P., Schirrmann, M., & Stolzenburg, F., (2023). Improving Deep Learning-based Plant Disease Classification with Attention Mechanism. Gesunde Pflanzen, 75(1), 49-59. https://doi.org/10.1007/s10343-022-00796-y
2. Chao, H., Fenhua, W., & Ran, Z., (2019). Sign language recognition based on cbam-resnet. Proceedings of the 2019 International Conference on Artificial Intelligence and Advanced Manufacturing, Ireland, 48, 1-6. https://doi.org/10.1145/3358331.3358379
3. DeGeorge, K.C, Holt, H.R., & Hodges, S.C. (2017). Bladder cancer: diagnosis and treatment. American family physician, 96(8): 507-514.
4. Du, Y., Yang, R., Chen, Z., Wang, L., Weng, X., & Liu, X., (2021). A deep learning network‐assisted bladder tumour recognition under cystoscopy based on Caffe deep learning framework and EasyDL platform. The International Journal of Medical Robotics and Computer Assisted Surgery, 17(1), 1-8. https://doi.org/10.1002/rcs.2169
5. Gao, F., Wu, T., Li, J., Zheng, B., Ruan, L., Shang, D., & Patel, B., (2018). SD-CNN: A shallow-deep CNN for improved breast cancer diagnosis. Computerized Medical Imaging and Graphics, 70, 53-62. https://doi.org/10.1016/j.compmedimag.2018.09.004
6. Ikeda, A., Nosato, H., Kochi, Y., Kojima, T., Kawai, K., Sakanashi, H., Murakawa, M., & Nishiyama., H., (2020). Support system of cystoscopic diagnosis for bladder cancer based on artificial intelligence. Journal of endourology, 34(3),352-358. https://doi.org/10.1089/end.2019.0509
7. Inoue, K., Fukuhara, H., Shimamoto, T., Kamada, M., Iiyama, T., Miyamura, M., Kurabayashi, A., Furihata, M., Tanimura, M., & Watanabe, H., (2012). Comparison between intravesical and oral administration of 5‐aminolevulinic acid in the clinical benefit of photodynamic diagnosis for nonmuscle invasive bladder cancer. Cancer, 118(4), 1062-1074. https://doi.org/10.1002/cncr.26378
8. Langlotz, C.P., Allen, B., Erickson, B.J., Kalpathy-Cramer, J., Bigelow, K., Cook, T.S., Flanders, A.E., Lungren, M.P., Mendelson, D.S., & Rudie, J.D., (2019). A roadmap for foundational research on artificial intelligence in medical imaging: from the 2018 NIH/RSNA/ACR/The Academy Workshop. Radiology, 291(3), 781-791. https://doi.org/10.1148/radiol.2019190613

9. Lazo, J.F., Moccia, S., Marzullo, A., Catellani, M., De Cobelli, O., Rosa, B., de Mathelin, M., & De Momi, E., (2021). A transfer-learning approach for lesion detection in endoscopic images from the urinary tract. arXiv preprint arXiv:210403927.
10. Lazo, J.F., Rosa, B., Catellani, M., Fontana, M., Mistretta, F.A., Musi, G., de Cobelli, O., de Mathelin, M., & De Momi, E., (2023). Semi-supervised Bladder Tissue Classification in Multi-Domain Endoscopic Images. IEEE Transactions on Biomedical Engineering, 70(10), 2822-2833. https://doi.org/10.1109/TBME.2023.3265679
11. Mateen, M., Wen, J., Nasrullah, Song, S., & Huang Z., (2018). Fundus image classification using VGG-19 architecture with PCA and SVD. Symmetry, 11(1), 1. https://doi.org/10.3390/sym11010001
12. Mukhtorov, D., Rakhmonova, M., Muksimova, S., & Cho, Y-I., (2023). Endoscopic image classification based on explainable deep learning. Sensors, 23(6), 3176. https://doi.org/10.3390/s23063176
13. Sertkaya, M.E., Ergen, B., & Togacar, M., (2019). Diagnosis of eye retinal diseases based on convolutional neural networks using optical coherence images. 2019 23rd International conference electronics, Lithuania, 1-5. https://doi.org/10.1109/ELECTRONICS.2019.8765579
14. Shirasuna, V.Y., & Gradvohl A., (2023). An optimized training approach for meteor detection with an attention mechanism to improve robustness on limited data. Astronomy and Computing, 45, 100753. https://doi.org/10.1016/j.ascom.2023.100753
15. Shkolyar, E., Jia, X., Chang, T.C., Trivedi, D., Mach, K.E., Meng, MQ-H., Xing, L., & Liao, J.C., (2019). Augmented bladder tumor detection using deep learning. European urology, 76(6), 714-718. https://doi.org/10.1016/j.eururo.2019.08.032
16. Siegel, R.L., Miller, K.D., Fuchs, H.E., & Jemal, A., (2021). Cancer statistics, 2021. Ca Cancer J Clin, 71(1), 7-33.
17. Toğaçar, M., Özkurt, K.B., Ergen, B., & Cömert, Z., (2020). BreastNet: A novel convolutional neural network model through histopathological images for the diagnosis of breast cancer. Physica A: Statistical Mechanics and its Applications, 545, 123592. https://doi.org/10.1016/j.physa.2019.123592
18. Wang, D., Gao, F., Dong, J., & Wang, S., (2019). Change detection in synthetic aperture radar images based on convolutional block attention module. 2019 10th International Workshop on the Analysis of Multitemporal Remote Sensing Images (MultiTemp), China, 1-4. https://doi.org/10.1109/Multi-Temp.2019.8866962
19. Woo S., Park, J., Lee J-Y., & Kweon, I.S., (2018). Cbam: Convolutional block attention module. Proceedings of the European conference on computer vision (ECCV), 3-19. https://doi.org/10.48550/arXiv.1807.06521
20. Yang, R., Du, Y., Weng, X., Chen, Z., Wang, S., & Liu, X., (2021). Automatic recognition of bladder tumours using deep learning technology and its clinical application. The International Journal of Medical Robotics and Computer Assisted Surgery, 17(2), e2194. https://doi.org/10.1002/rcs.2194
21. Zhou, F., Liu, X., Zhang, X., Liu, Y., Jia, C., & Wu, C., (2022). Keyhole status prediction based on voting ensemble convolutional neural networks and visualization by Grad-CAM in PAW. Journal of Manufacturing Processes, 80,805-815. https://doi.org/10.1016/j.jmapro.2022.06.03