Comparative Analysis of Hyperparameter Tuned Convolutional Neural Networks for classification of Diabetic Retinopathy
Keywords:
Hyperparameters tuning, Machine learning, Diabetic RetinopathyAbstract
Diabetic Retinopathy (DR) is one of the leading cause for loss of vision among diabetic patients. Early detection and treatment of disease can prevent serious complications. Deep learning in medical image analysis is very effective. One of the important aspects in building efficient machine learning algorithm is to choose the right combination of hyper parameters. To understand the sensitivity of hyper parameter, hyperparameter tuning is applied to ten deep convolutional neural network (DCNN). The fundus images from Kaggle diabetic retinopathy dataset is pre-processed and resized to 224 x 224x3. Depending upon the severity of disease dataset images belongs to one of the class (0, 1, 2, 3, 4) so performance of all the ten networks are evaluated class wise. Experimental results reveal that Vgg16 outperformed InceptionV3, Xception, ResNet50, DenseNet121, Densenet169 and DenseNet201. MobileNetv2 outperformed other two light weight models MobileNetv1 and NASNetMobile. Training accuracy for vgg16 is 92%, validation accuracy is 85%, sensitivity is 84%, specificity is 96% and f1 score is 0.84. For MobileNetv2 training accuracy is 98%, validation accuracy is 81%, sensitivity is 81%, specificity is 94% and f1 score is 0.81. Training of Vgg16 and MobileNetv2 is carried out using Adam optimizer with learning rate of 0.001 and 0.00002 respectively, dropout 0.5, batch size is set to 32 and no. of Epoch to 40 and 20 respectively. While comparing the proposed work with previously related similar work it is found that our approach yields better results. This work contributes towards the improvement of the image classification techniques.
Downloads
References
Saeedi, Pouya, et al (2019) Global and regional diabe-tes prevalence estimates for 2019 and projections for 2030 and 2045: Results from the International Diabe-tes Federation Diabetes Atlas. Diabetes research and clinical practice 157, pp 107843.
Asiri, N. , et.al (2019) Deep learning based computer-aided diagnosis systems for diabetic retinopathy: A survey. Artificial intelligence in medicine, 99, pp 101701
Rupa Patel, Anita Chaware (2021). Computer Vision and Recognition Systems Using Machine and Deep Learning Approaches , In: The Institution of Engineer-ing and Technology ,pp 253-276
K. Simonyan, A. Zisserman (2014) Very deep convo-lutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.
K. He, X. Zhang et.al (2015) Deep residual learning for image recognition. arXiv preprint arXiv:1512.03385
C. Szegedy, W. Liu, et.al(2015) Going Deeper with Convolutions, In :The IEEE Conference on Computer Vision and Pattern Recognition
F. Chollet et. al (2016) Xception: Deep Learning with Depth wise Separable Convolutions
Mishra, Supriya; Hanchate, Seema; Saquib, Zia (2020). Diabetic Retinopathy Detection using Deep Learning,IEEE 2020 International Conference on Smart Technologies in Computing, Electrical and Elec-tronics , Bengaluru ,pp 515–520. doi:10.1109/ICSTCEE49637.2020.9277506
Huang, G., Liu, Z., Van Der Maaten, L. and Wein-berger, K.Q.( 2017) Densely connected convolutional networks. In Proceedings of the IEEE conference on computer vision and pattern recognition ,pp. 4700-4708. http://doi.org/10.1109/CVPR.2017.243
Howard, A. G., Zhu, M., Chen, et.al., “ Mobilenets: Efficient convolutional neural networks for mobile vi-sion applications,” arXiv preprint arXiv:1704.04861,2017.
Mark, S. et al., “Mobilenetv2: Inverted residuals and linear bottlenecks,” in Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2018, pp 4510-4520.
Zoph, B., Vasudevan, V., Shlens, J. and Le, Q.V.(2018) Learning transferable architectures for scalable image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition pp. 8697-8710.
Weerts , H.J., Mueller, A.C. and Vanschoren, J.(2020). Importance of tuning hyperparameters of machine learning algorithms. arXiv preprint arXiv:2007.07588.
Jacek M. Zurada (1994) Introduction to Artificial Neural System, Jaico Publishing House.
Jin H. et al.(2022) Hyperparameter Importance for Machine Learning Algorithms. arXiv preprint arXiv:2201.05132
A. Gulcu ,Z. KUş,(2020),Hyper-Parameter Selection in Convolutional Neural Networks Using Microcanonical Optimization Algorithm, in IEEE Access, vol. 8, pp. 52528-52540, doi: 10.1109/ACCESS.2020.2981141
Jafar A, Myungho L. (2020) Hyperparameter Optimi-zation for Deep Residual Learning in Image Classifi-cation. In2020 IEEE International Conference on Au-tonomic Computing and Self-Organizing Systems Companion (ACSOS-C) ,pp. 24-29
Yu T, Zhu H. (2020) Hyper-parameter optimization: A review of algorithms and applications. arXiv preprint arXiv:2003.05689.
Amirabadi, M.A., Kahaei, M.H. et al.( 2020) Novel suboptimal approaches for hyperparameter tuning of deep neural network [under the shelf of optical com-munication]. Physical Communication, 41, p.101057.
Lerman, P.M., (1980) Fitting segmented regression models by grid search. Journal of the Royal Statistical Society: Series C (Applied Statistics), 29(1), pp.77-84.
James Bergstra , Bengio (2012), Random search for hyper-parameter optimization. Journal of machine learning research, 13(2).
Bergstra, J., et al.(2011) Algorithms for hyper-parameter optimization. Advances in neural infor-mation processing systems, 24.
Loshchilov, I. and Hutter, F., (2016) CMA-ES for hy-perparameter optimization of deep neural net-works. arXiv preprint arXiv:1604.07269.
Mockus, J., (2012). Bayesian approach to global op-timization: theory and applications (Vol. 37). Springer Science & Business Media.
Pratt, H., Coenen, F., Broadbent, D.M., Harding, S.P. and Zheng, Y., 2016. Convolutional neural networks for diabetic retinopathy. Procedia computer sci-ence, 90, pp.200-205
Pradhan A, Dey BK, Nath RK, Sarma B.(2020) Trans-fer learning based classification of diabetic retinopa-thy stages. In2020 International Conference on Com-putational Performance Evaluation (ComPE),IEEE ,pp. 813-819.
Vimukthi, Y., Kodikara, N. and Nanayakkara, L., (2020). A Deep Learning Based Approach for the Classification of Diabetic Retinopathy in Human Ret-ina.
Mohammadian, S., Karsaz, A. and Roshan, Y.M., 2017, November. Comparative study of fine-tuning of pre-trained convolutional neural networks for diabetic retinopathy screening. In 2017 24th National and 2nd International Iranian Conference on Biomedical En-gineering (ICBME) (pp. 1-6). IEEE.
https://www.kaggle.com/c/aptos2019-blindness-detection/data
Arora, M. and Pandey, M., (2019), February. Deep neural network for diabetic retinopathy detection. In 2019 International Conference on Machine Learn-ing, Big Data, Cloud and Parallel Computing (COMITCon) ,pp. 189-193, IEEE.
Rupa Patel and Anita Chaware,(2020),Transfer Learning with Fine-Tuned MobileNetV2 for Diabetic Retinopathy, In 2020 International Conference for Emerging Technology (INCET) , Belgaum, India, pp. 1-4, IEEE. doi: 10.1109/INCET49848.2020.9154014
Rupa Patel and Anita Chaware (2021) Quantizing MobileNet Models for Classification Problem “,INDIACom-2021; 15th INDIACom; 2021 8th In-ternational Conference on Computing for Sustainable Global Development,organised by BVICAM, New Delhi.
Asiri, N., Hussain, M., Al Adel, F. and Alzaidi, N., (2019.) Deep learning based computer-aided diagnosis systems for diabetic retinopathy: A survey. Artificial intelligence in medicine, 99, p.101701
Wang YE, Wei GY, Brooks D.,(2019) Benchmarking tpu, gpu, and cpu platforms for deep learning. arXiv preprint arXiv:1907.10701.
Rawat, Waseem & Wang, Zenghui. (2017). Deep Convolutional Neural Networks for Image Classifica-tion: A Comprehensive Review. Neural Computation. 29. 1-98. 10.1162/NECO_a_00990
Aswathi, T., Swapna, T.R. and Padmavathi, S., (2021), Transfer Learning approach for grading of Diabetic Retinopathy. In Journal of Physics: Conference Se-ries (Vol. 1767, No. 1, p. 012033). IOP Publishing.
Rambabu, B. ., Vikranth, B. ., Anupkanth, S. ., Samya, B. ., & Satyanarayana, N. . (2023). Spread Spectrum based QoS aware Energy Efficient Clustering Algo-rithm for Wireless Sensor Networks . International Journal on Recent and Innovation Trends in Compu-ting and Communication, 11(1), 154–160. https://doi.org/10.17762/ijritcc.v11i1.6085
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.
