Design of intelligent lesion detection system using deep architecture neural networks in the lower limb X-ray images
Diagnosis of musculoskeletal abnormalities is very important because more than 1.7 billion people worldwide are affected by musculoskeletal disorders. In this study, we focus on the diagnosis of musculoskeletal abnormalities in the lower extremities in X-ray images. The interpretation of these images is the responsibility of an experienced physician and radiologist. The physician's experience has a major impact on the accuracy of the diagnosis and can sometimes be very challenging for young physicians. We proposed a new deep neural network architecture with two different scenarios that perform the function of lower extremity lesion diagnosis with great accuracy. The core of the proposed method is a deep learning framework based on the Mask R-CNN and CNN. The model with the best results was the use of the Mask R-CNN algorithm to generate the bounding box and then the CNN algorithm to detect the class based on the bounding box. Our dataset contains 61,098 musculoskeletal radiographic images, which includes 42658 normal images and 18440 abnormal images. Each image belongs to one type of lower extremity radiography including the toe, foot, ankle, leg, knee, femur, and hip joint. The proposed model can detect different types of lower limb lesions. Our model achieves an AUC-ROC of 0.925, with an operating point of 0.859 sensitivity and 0.893 specificity. By comparing the different results, it can be concluded that the consecutive implementation of Mask R-CNN and CNN works better than Mask R-CNN and CNN separately.