Ultrasonic thyroid nodule detection method based on U-Net network

Chu, Cong Qiu; Zheng, Jihui; Yong, Zhou

doi:10.1016/j.cmpb.2020.105906

Cited by 36 publications

(18 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Driven by the continuous development of deep learning algorithm and graphics processing unit (GPU), CAD has also entered a new research stage. In the recent years, the deep learning technology has been applied in multiple fields which involve the CAD systems, focusing on the medical image analysis of computed tomography (CT) [19,20], magnetic resonance imaging (MRI) [21,22], X-ray [23], ultrasonic examination [24], endoscope [25], and pathological slides [26][27][28][29].…”

Section: Computer-assisted Diagnosis Via Deep Learning Methodsmentioning

confidence: 99%

NECScanNet: Novel Method for Cervical Neuroendocrine Cancer Screening from Whole Slide Images

Liao

Qin

Zheng

2021

Security and Communication Networks

View full text Add to dashboard Cite

As a rare malignant tumor, cervical neuroendocrine cancer (NEC) is difficult in diagnosis even for experienced pathologists. A computer-assisted diagnosis may be helpful for the improvement of diagnostic accuracy. Nevertheless, the computer-aided pathological diagnosis has to face a great challenge that the hundred-million-pixels or even gig-pixels whole slide images (WSIs) cannot be applied directly in the existing deep convolution network for training and analysis. Therefore, the construction of a neural network to realize the automatic screening of cervical NEC is challenging; meanwhile, as far as we know, little attention has been paid to this field. In order to address this problem, here we present a multiple-instance learning method for automatic recognition of cervical NEC on pathological WSI, which consists of the Sliding Detector module and Lesion Analyzer module. A pathological WSI dataset, which is composed of 84 NEC cases and 216 NEC-free cases from the Pathological Department of West China Second University Hospital, is applied to evaluate the performance of the method. The experimental results show that the recall rate, accuracy rate, and precision rate of our method for automatic recognition are 92.9%, 92.7%, and 83.0%, respectively, demonstrating the effectiveness and the potential in clinical practice. The application of this method in computer-assisted pathological diagnosis is expected to decrease the misdiagnosis as well as the false diagnosis of rare cervical NEC, and, consequently, improve the therapeutic effect of cervical cancers.

show abstract

Section: Computer-assisted Diagnosis Via Deep Learning Methodsmentioning

confidence: 99%

NECScanNet: Novel Method for Cervical Neuroendocrine Cancer Screening from Whole Slide Images

Liao

Qin

Zheng

2021

Security and Communication Networks

View full text Add to dashboard Cite

show abstract

“…We compare with SOTA segmentation methods proposed for the three specific tasks and general medical image seg- I. In the table, ResUnet [11] and BEAL [18] are proposed for optic disc/cup segmentation, TransBTS [3] and EnsemDiff [19] are proposed for the brain tumor segmentation, MTSeg [20] and UltraUNet [21] are proposed for the Thyroid Nodule segmentation, CENet [22], MRNet [1], SegNet [23], nnUNet [24] and TransUNet [2] are proposed for the general medical image segmentation. We evaluate the segmentation performance by Dice score and IoU.…”

Section: Resultsmentioning

confidence: 99%

MedSegDiff: Medical Image Segmentation with Diffusion Probabilistic Model

Wu¹,

Fang²,

Yang³

et al. 2022

Preprint

View full text Add to dashboard Cite

Diffusion probabilistic model (DPM) recently becomes one of the hottest topic in computer vision. Its image generation application such as Imagen, Latent Diffusion Models and Stable Diffusion have shown impressive generation capabilities, which aroused extensive discussion in the community. Many recent studies also found it is useful in many other vision tasks, like image deblurring, super-resolution and anomaly detection. Inspired by the success of DPM, we propose the first DPM based model toward general medical image segmentation tasks, which we named MedSegDiff. In order to enhance the step-wise regional attention in DPM for the medical image segmentation, we propose dynamic conditional encoding, which establishes the state-adaptive conditions for each sampling step. We further propose Feature Frequency Parser (FF-Parser), to eliminate the negative effect of high-frequency noise component in this process. We verify MedSegDiff on three medical segmentation tasks with different image modalities, which are optic cup segmentation over fundus images, brain tumor segmentation over MRI images and thyroid nodule segmentation over ultrasound images. The experimental results show that MedSegDiff outperforms stateof-the-art (SOTA) methods with considerable performance gap, indicating the generalization and effectiveness of the proposed model.

show abstract

“…The segmentation methods commonly employed in thyroid nodules are mainly divided into the following categories: active contour [3], feature extraction and classification [4], and deep learning [5], etc. Although the traditional segmentation methods have made some progress in CAD, the feature selection and parameter setting have more subjective factors.…”

Section: Segmentation Methodsmentioning

confidence: 99%

“…Recently, convolutional neural networks (CNNs) have shown desirable results in the segmentation task of medical ultrasound images. Chen et al proposed a UNet bases marker-guided segmentation model of thyroid nodules, with a segmentation accuracy of 98% [5].…”

Section: Segmentation Methodsmentioning

confidence: 99%

A Deep Learning-Based Cascade Automatic Classification System for Malignant Thyroid Nodule Recognition in Ultrasound Image

Zhou

Wang

Zhou

et al. 2022

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

As the most common malignancy in the endocrine system, thyroid cancer is often determined based on ultrasound images by professional radiologists based on ultrasound images. In this study, a cascade automatic classification system (CACS) is proposed to accurately achieve automatic segmentation and classification of thyroid nodules, reducing over-reliance on experienced doctors in the diagnosis. Firstly, the Res-UNet is used to segment the region of interest (ROI) of thyroid nodules in ultrasound images. Then, to overcome the limitations of the small sample size of medical training data as well as category imbalance, the modified Conditional Variational Auto-Encoder (CVAE), learning the probability distribution of the ultrasound images, is trained to produce new medical ultrasound images for ROI data augmentation. Finally, we perform the classification task by a three-branch ensemble network combined with clinical diagnostic features of thyroid nodule such as the aspect ratio, internal and external echoes, and edges. The proposed CACS can identify benign and malignant thyroid nodules with an accuracy of 0.892 and an AUC of 0.887, according to testing on real patient data.

show abstract

Ultrasonic thyroid nodule detection method based on U-Net network

Cited by 36 publications

References 20 publications

NECScanNet: Novel Method for Cervical Neuroendocrine Cancer Screening from Whole Slide Images

NECScanNet: Novel Method for Cervical Neuroendocrine Cancer Screening from Whole Slide Images

MedSegDiff: Medical Image Segmentation with Diffusion Probabilistic Model

A Deep Learning-Based Cascade Automatic Classification System for Malignant Thyroid Nodule Recognition in Ultrasound Image

Contact Info

Product

Resources

About