Automatic body region localization in 3D-CT images based on the improved YOLO model

Linghao Du,Qingyi Liu,Yande Ren,Rui Wang,Lin Cui,Xiaolin Min,Kai Huang,Peirui Bai,J Zhu

doi:10.1051/matecconf/202235503022

Abstract

Automatic body region localization in medical three-dimensional (3D)-CT images is a critical step of computerized body-wide Automatic Anatomy Recognition (AAR) system, which can be applied for radiotherapy planning and interest slices retrieving. Currently, the complex internal structure of human body and time consuming computation are the main challenges for the localization. Therefore, this paper introduces and improves the YOLO-v3 model into the body region localization for these problems. First, seven categories of body regions in a CT volume image I are defined based on the modification version of our previous work. Second, an improved YOLO-v3 model is trained to classify each axial slice into one of the seven categories. Then, the effectiveness of the proposed method is evaluated on 3D-CT images that collected from 220 subjects. The experimental results demonstrate that the slice localizing error is less than 3 NoS (Number of slices), which is competitive to the state-of-the-art methods. Beyond this, our method is simple and computationally efficient owing to its less training time, and the average computational time for localizing a volume CT images is about 3 second, which shows potential for a further application.

Highlights

1.1 Background and related workAutomatic body region localization based on three-dimensional medical Computed Tomography (3D-CT) is a critical step of Anatomy Recognition (AAR), which can be applied for radiotherapy planning and interest slices retrieving etc [1]
Automatic body region localization based on three-dimensional medical Computed Tomography (3D-CT) is a critical step of AAR, which can be applied for radiotherapy planning and interest slices retrieving etc [1]
Let I represent a scan or image, the location of the superior axial slice of the thorax in I was denoted by TS(I), and the location of its inferior axial slice was denoted by TI(I), as listed in table 1[7]

Summary

Introduction

1.1 Background and related workAutomatic body region localization based on three-dimensional medical Computed Tomography (3D-CT) is a critical step of AAR, which can be applied for radiotherapy planning and interest slices retrieving etc [1]. The earliest body region localizing method of CT image is to manually extract the scanning and spacing parameters from the header file of DICOM format (Digital Imaging and Communications in Medicine). Hong et al invented a device and algorithm for locating the neck, thorax, abdomen and pelvis sequentially in terms of the labeled reference frame of the whole body[5] Both of the localizing methods are applicable to multimodal medical image data. It was difficult to discern them automatically by computer

Methods

Results

Conclusion