Exploring the impact of hyperparameter and data augmentation in YOLO V10 for accurate bone fracture detection from X-ray images

Abstract

Accurately identifying bone fractures from the X-ray image is essential to prompt timely and appropriate medical treatment. This research explores the impact of hyperparameters and data augmentation techniques on the performance of the You Only Look Once (YOLO) V10 architecture for bone fracture detection. While YOLO architectures have been widely employed in object detection tasks, recognizing bone fractures, which can appear as subtle and complicated patterns in X-ray images, requires rigorous model tuning. Image augmentation was done using the image unsharp masking approach and contrast-limited adaptive histogram equalization before training the model. The augmented images assist in feature identification and contribute to overall performance of the model. The current study has performed extensive experiments to analyze the influence of hyperparameters like the number of epochs and the learning rate, along with the analysis of the data augmentation on the input data. The experimental outcome has proven that particular hyperparameter combinations, when paired with targeted augmentation strategies, improve the accuracy and precision of fracture detection. It is observed that the proposed model yielded an accuracy of 0.964 on evaluation over the augmented data. The statistical analysis of the classification precision across the augmented and raw images is observed as 0.98 and 0.95, respectively. In comparison with other deep learning models, the empirical evaluation of the YOLO V10 model clearly demonstrates its superior performance over conventional approaches for bone fracture detection.

Keywords: Bone fracture; Data augmentation; Deep learning; Hyperparameters; YOLO model.

Fig. 1

The block diagram of the YOLO V10-based bone fracture detection module.

Fig. 2

A summary of various versions of the YOLO model is provided.

Fig. 3

The workflow of the YOLO V10 model for bone fracture detection.

Fig. 4

Graph representing the distribution of classes in the dataset.

Fig. 5

Sample images that are obtained from the dataset.

Fig. 6

Dataset Samples and number of instances used in data augmentation.

Fig. 7

The input and output images and histogram values of image unsharp masking approach.

Fig. 8

The input and output images and histogram values of image unsharp masking approach.

Fig. 9

The architecture diagram of CSPNet.

Fig. 10

The architecture diagram of the Path Aggregation Network.

Fig. 11

The architecture diagram of the head component of the YOLO V10 model.

Fig. 12

The training and validation loss metrics were obtained on standard hyperparameters.

Fig. 13

The training and validation loss metrics were obtained over configuration #1.

Fig. 14

The training and validation loss metrics were obtained over configuration #2.

Fig. 15

The training and validation loss metrics were obtained over configuration #3.

Fig. 16

The training and validation loss metrics were obtained over configuration #4.

Fig. 17

The training and validation loss metrics were obtained on standard hyperparameters without data augmentation.

Fig. 18

The confusion matrix on experimenting over the YOLO V10, (a) Augmented images, (b) raw X-Ray images.

Fig. 19

The experimental outcome of the YOLO V10 model on augmented X-ray images.

Fig. 20

The experimental outcome of the YOLO V10 model on raw X-ray images.