Citation

Murugan, Pallikonda Rajasekaran and Thangaperumal, Sivashankari and Hossen, Md. Jakir and Wong, Wai Kit and Ng, Poh Kiat and Ramaraj, Kottaimalai and Thiyagarajan, Arunprasath and Kamath, Rajesh (2025) Enhancement of Anterior Cruciate Ligament Tear Diagnosis using Optimized Hyperparameters in Deep Learning Models. In: 2025 Multimedia University Engineering Conference, MECON 2025, 21 July 2025 - 23 July 2025, Cyberjaya, Malaysia.

Text 71.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

Musculoskeletal disorders must be treated early, and an accurate diagnosis is very important. Among them Anterior Crucial Ligament (ACL) tears occur mostly in athletes (50%-70%), aged people (5%-10%) and people with previous injuries (20%-30%). ACL tears are complicated to identify, and segmentation is also a tedious process because of their complex morphological structures. Along with manual interpretation, Magnetic Resonance Imaging (MRI) diagnosis is essential in confirming the ACL tears. An automated decision support system aids medical professionals in detecting and confirming the ACL tears. Convolutional Neural Network (CNN) models are efficient in feature selection as it recognizes textures with local connectivity and spatial hierarchy. Hyperparameter values vary depending on the dataset and neural network architectures. As manual tuning is performed unsystematically, it is time-consuming to identify the correct parameters. So, the proposed work trained CNN-based deep learning models to classify knee MRI into normal, partial torn and full torn ACL. Also, the hyperparameters are optimized efficiently by Bayesian optimization techniques. With the implementation of transfer learning, SqueezeNet, DarkNet-53 and DenseNet-201 are compared with the performance of optimized hyperparameters. After Bayesian Optimization, these models performed well with greater accuracy and less computational time. In conclusion, optimized SqueezeNet, DarkNet-53 and DenseNet-201 architectures provide an efficient decision support system to assist healthcare professionals in diagnosis of ACL tears.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Diagnosis, ACL tears
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science R Medicine > R Medicine (General) > R858-859.7 Computer applications to medicine. Medical informatics
Divisions:	Faculty of Engineering and Technology (FET)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	18 Mar 2026 07:40
Last Modified:	18 Mar 2026 07:40
URII:	http://shdl.mmu.edu.my/id/eprint/15560

Downloads

Edit (login required)