Comparative Analysis of ResNet101, InceptionV3, and InceptionResnetV2 Architectures for Cervical Vertebrae Maturation Stage Classification

The classification of cervical vertebrae maturity (CVM) plays a crucial role in orthodontic diagnosis and treatment planning. This study aims to evaluate the performances of bone maturity classification using three convolutional neural network architectures. The dataset used in this study utilizes the CVM-900 dataset which contains labeled cervical vertebrae x-ray images in 6 levels of bone maturation. Preprocessing was conducted by ROI cropping which results in three datasets: Crop C2-C4, Crop C2-C6, and uncrop. Then, to decrease the possibility of overfitting, the datasets are augmented with random translation and rotation. This work compares three pre-trained networks: ResNet-101, InceptionResNetV2, and InceptionV3. Each network has unique characteristics to improve performance in deep learning. Visual explanations of CVM stage classification from each model are conducted through Grad-CAM which leverages the gradient information to the target classes. We found that InceptionResNetV2 yields the highest accuracy across models and ROIs. It is shown that ROI affects model's performance. However, due to overfitting and the inability of our model to infer multiscale features, the increase in accuracy is not as significant.

I. Introduction

Bone age measurement is an important examination in the medical field. Bone age is an indicator of a person's level of bone and biological maturity [1]. Bone age has a higher degree of accuracy for defining a person's biological age compared to birth age. In addition, bone age is closely related to the level of bone maturity. The level of bone maturity is widely used to determine medical procedures. Some of the important uses of knowing the level of bone maturity are determining the general biological growth rate in children, predicting the development of children with certain diseases such as down syndrome [2], and determining the type of treatment for patients with endocrine problems such as osteoporosis [3]. Bone maturity level is also information that is often used by dentists. One of the urgent ways of determining the level of bone maturity is to determine the treatment plan for children with cleft lips. Children with unilateral cleft lip are known to have slower bone growth compared to children with normal conditions [4]. Treatment errors in bone conditions that do not have the appropriate level of maturity can be fatal.