Hyper-tuned convolutional neural network based pediatric skeletal bone age estimation model
DOI:
https://doi.org/10.18203/2320-6012.ijrms20260255Keywords:
CNN, Deep Learning, Hyper-parameter, Metaheuristic, Pediatric skeletal bone age, Walrus optimization, RSNAAbstract
Background: The pediatric skeletal bone age estimation model is developed to estimate the bone age in order to investigate the genetic and growth disorders of children’s. In this paper, a deep learning-based is presented for it.
Method: Initially, in this model, the standard RSNA dataset of hand X-ray images is read. Followed by median filtering to remove noise and segmentation of the hand bone region using k-mean clustering. After that, the Convolutional Neural Network (CNN) algorithm is used to estimate the bone age. In this article, the CNN algorithm is used over other deep learning algorithms due to its automatically extracting the features from the hand X-ray images and estimating the bone age. Furthermore, in this research, the hyper-parameter optimization of the CNN algorithm is done by finding the best parameter values using the metaheuristic algorithm to enhance the performance of it. The metaheuristic walrus optimization algorithm is used, and it determines the general hyperparameters, such as the learning rate of the CNN algorithm, based on the objective function.
Results: The simulation evaluation was done on MATLAB 2018b software. The standard RSNA dataset of hand X-ray images was used. The performance evaluation is done by splitting the same dataset into training and testing ratios and evaluated using the error metrics. The result indicates that the proposed model accomplishes the lower values of these error metrics over the previous approaches.
Conclusion: The proposed method efficiently measures pediatric skeletal bone age by processing hand X-ray images with the CNN algorithm, which has been optimized through hyperparameter tuning.
Metrics
References
Liu C, Xie H, Liu Y, Zha Z, Lin F, Zhang Y. Extract bone parts without human prior: End-to-end convolutional neural network for pediatric bone age assessment. In International conference on medical image computing and computer-assisted intervention. 2019: 667-675. DOI: https://doi.org/10.1007/978-3-030-32226-7_74
Jian K, Li S, Yang M, Wang S, Song C. Multi-characteristic reinforcement of horizontally integrated TENet based on wrist bone development criteria for pediatric bone age assessment. Applied Intell. 2023;53(19):22743-52. DOI: https://doi.org/10.1007/s10489-023-04633-1
Jabbar AJ, Abdulmunem AA. Bone age assessment using deep learning architecture: A Survey. Int Conf Intelligent Syst Comp Vision. 2022;3:1-6. DOI: https://doi.org/10.1109/ISCV54655.2022.9806110
Alzubaidi L, Khamael AD, Salhi A, Alammar Z, Fadhel MA. Comprehensive review of deep learning in orthopaedics: Applications, challenges, trustworthiness, and fusion. Artificial Int Med. 2024;155:102935. DOI: https://doi.org/10.1016/j.artmed.2024.102935
Bram JT, Pareek A, Beber SA, Jones RH, Shariatnia MM. Determination of Skeletal Age From Hand Radiographs Using Deep Learning. The American J Sports Med. 2025;53(11):2715-25. DOI: https://doi.org/10.1177/03635465251359618
Li S, Liu B, Li S, Zhu X, Yan Y, Zhang D. A deep learning-based computer-aided diagnosis method of X-ray images for bone age assessment. Complex Intell Sys. 2022;8(3):1929-39. DOI: https://doi.org/10.1007/s40747-021-00376-z
Liu ZQ, Hu ZJ, Wu TQ, Ye GX, Tang YL, Zeng ZH. Bone age recognition based on mask R-CNN using xception regression model. Fronti Physiol. 2023;14:1062034. DOI: https://doi.org/10.3389/fphys.2023.1062034
Wang S, Jin S, Xu K, She J, Fan J, He M. A pediatric bone age assessment method for hand bone X-ray images based on dual-path network. Neural Comp Appl. 2024;36(17):9737-52. DOI: https://doi.org/10.1007/s00521-023-09098-4
Kim KD, Kyung S, Jang M, Ji S, Lee DH, Yoon HM, Kim N. Enhancement of Non-Linear Deep Learning Model by Adjusting Confounding Variables for Bone Age Estimation in Pediatric Hand X-rays. J Dig Imag. 2023;36(5):2003-14. DOI: https://doi.org/10.1007/s10278-023-00849-2
Nivedita, Solanki S. Enhancing the Accuracy of Automatic Bone Age Estimation Using Optimized CNN Model on X-Ray Images. InInternational Conference on Machine Learning Algorithms 2024: 329-340. DOI: https://doi.org/10.1007/978-3-031-75861-4_29
Saadi M, Aljobouri HK, Al-Waely NK. Improving Bone Age Assessment with Inception-V3 and Faster R-CNN. InInnovative and Intelligent Digital Technologies. Inc Effi. 2024;1:579-90. DOI: https://doi.org/10.1007/978-3-031-70399-7_44
Liu B, Zhang Y, Chu M, Bai X, Zhou F. Bone age assessment based on rank-monotonicity enhanced ranking CNN. IEEE. 2019;7:120976-83. DOI: https://doi.org/10.1109/ACCESS.2019.2937341
Reddy NE, Rayan JC, Annapragada AV, Mahmood NF. Bone age determination using only the index finger: a novel approach using a convolutional neural network compared with human radiologists. Ped Radiol. 2020;50(4):516-23. DOI: https://doi.org/10.1007/s00247-019-04587-y
Wojciuk M, Swiderska-Chadaj Z, Siwek K, Gertych A. Improving classification accuracy of fine-tuned CNN models: Impact of hyperparameter optimization. Heliyon. 2024;10(5):856. DOI: https://doi.org/10.1016/j.heliyon.2024.e26586
Tomar V, Bansal M, Singh P. Metaheuristic algorithms for optimization: A brief review. Engineering Proceedings. 2024;59(1):238. DOI: https://doi.org/10.3390/engproc2023059238
Raiaan MA, Sakib S, Fahad NM, Al Mamun A, Rahman MA, Shatabda S, Mukta MS. A systematic review of hyperparameter optimization techniques in Convolutional Neural Networks. Dec Anal J. 2024;11:100470. DOI: https://doi.org/10.1016/j.dajour.2024.100470
Pei N, Zhuang Y, Su Z, Wang F. Automated Bone Age Assessment and Adult Height Prediction from Pediatric Hand Radiographs via a Cascaded Deep Learning Framework. J Med Sys. 2025;49(1):170. DOI: https://doi.org/10.1007/s10916-025-02306-9
RSNA Pediatric Bone Age Challenge. RSNA. 2017. Available at: https://www.rsna.org/artificial-intelligence/ai-image-challenge/rsna-pediatric-bone-age-challenge. Accessed on 12 July 2025.
Rajitha B, Agarwal S. Segmentation of Epiphysis Region-of-Interest (EROI) using texture analysis and clustering method for hand bone age assessment. Multimedia Tools and Applications. 2022;81(1):1029-54. DOI: https://doi.org/10.1007/s11042-021-11531-6
Kasani AA, Sajedi H. Hand bone age estimation using divide and conquer strategy and lightweight convolutional neural networks. Engineering Appl Artif Intelligence. 2023;120:105935. DOI: https://doi.org/10.1016/j.engappai.2023.105935
Han M, Du Z, Yuen KF, Zhu H, Li Y, Yuan Q. Walrus optimizer: A novel nature-inspired metaheuristic algorithm. Expert Syst Appl. 2024;239:122413. DOI: https://doi.org/10.1016/j.eswa.2023.122413
Tan M, Le Q. Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning. 2019: 6105-6114.
Szegedy C, Vanhoucke V, Ioffe S, Shlens J, Wojna Z. Rethinking the inception architecture for computer vision. InProceedings of the IEEE conference on computer vision and pattern recognition 2016: 2818-2826. DOI: https://doi.org/10.1109/CVPR.2016.308
He K, Zhang X, Ren S, Sun J. Deep residual learning for image recognition. InProceedings of the IEEE conference on computer vision and pattern recognition. 2016: 770-778. DOI: https://doi.org/10.1109/CVPR.2016.90
Huang G, Liu Z, Van Der Maaten L, Weinberger KQ. Densely connected convolutional networks. InProceedings of the IEEE conference on computer vision and pattern recognition 2017: 4700-4708. DOI: https://doi.org/10.1109/CVPR.2017.243
Chollet F. Xception: Deep learning with depthwise separable convolutions. InProceedings of the IEEE conference on computer vision and pattern recognition 2017: 1251-1258. DOI: https://doi.org/10.1109/CVPR.2017.195
Downloads
Published
How to Cite
Issue
Section
