Texture Synthesis Based Thyroid Nodule Detection From Medical Ultrasound Images: Interpreting and Suppressing the Adversarial Effect of In-place Manual Annotation.

Siqiong Yao,Junchi Yan,Hui Lu,Biyun Qian,Mingyu Wu,Weituo Zhang,Xue Yang

doi:10.3389/fbioe.2020.00599

Siqiong Yao, Junchi Yan + Show 5 more

Open Access

https://doi.org/10.3389/fbioe.2020.00599

Copy DOI

Abstract

Deep learning method have been offering promising solutions for medical image processing, but failing to understand what features in the input image are captured and whether certain artifacts are mistakenly included in the model, thus create crucial problems in generalizability of the model. We targeted a common issue of this kind caused by manual annotations appeared in medical image. These annotations are usually made by the doctors at the spot of medical interest and have adversarial effect on many computer vision AI tasks. We developed an inpainting algorithm to remove the annotations and recover the original images. Besides we applied variational information bottleneck method in order to filter out the unwanted features and enhance the robustness of the model. Our impaiting algorithm is extensively tested in object detection in thyroid ultrasound image data. The mAP (mean average precision, with IoU = 0.3) is 27% without the annotation removal. The mAP is 83% if manually removed the annotations using Photoshop and is enhanced to 90% using our inpainting algorithm. Our work can be utilized in future development and evaluation of artificial intelligence models based on medical images with defects.

Highlights

Medical imaging is one of the most important sources of clinical diagnosis and clinical practice has accumulated and is currently generating huge amount of high-resolution medical imaging
We address the problem of computational and deep learning based thyroid nodule detection
Further improvement is made in a very recent work (Yu J. et al, 2018) whereby the boundary artifacts are effectively addressed by introducing a new context-aware layer, which can compensate the discontinuity between the surrounding content and the inpainting part

Summary

Introduction

Medical imaging is one of the most important sources of clinical diagnosis and clinical practice has accumulated and is currently generating huge amount of high-resolution medical imaging. It’s worth mentioning that ultrasonic inspection (UI) has some particular advantages in practice, by dynamic multi-trials, the area of interest can be effectively identified, leading to better (qualitative) measurement of size, quantity, diolame, calcification, and the relationship between different tissues of organizations. All such information is critical and helpful to the early screening and guidance before and after surgery and other medical treatment. The application of deep learning in ultrasound images is expected to be used in some areas like identification of benign and malignant nodules or the nodules detection

Objectives

Methods

Results