DBT Masses Automatic Segmentation Using U-Net Neural Networks

Lai, Xiaobo; Yang, W.; Li, Ruipeng

doi:10.1155/2020/7156165

Cited by 18 publications

(11 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Later on, clinicians from "Hospital 20 de Noviembre" in Mexico city provided 120 sequences with the four chambers segmented, so that the systems could be retrained once again and were capable of localising all four chambers of the heart. H&E WSI Predicts origins for cancers of unknown primary TCGA TOAD 18 [117] H&E WSI Identify the sub-type of renal cell carcinoma TCGA, CPTAC , CAMELYON16 and CAMELYON17 CLAM 3 [118] H&E WSI and WBCs Nuclei instance segmentation GlaS, GRAG, MonuSeg, CPM, and WBC NuClick 7 [59] Breast image analysis Mammograms Breast mass segmentation and shape classification DDSM and REUS Hospital cGAN and CNN 4 [119] MRI Breast tumors classification BI-RADS Pre-trained CNNs 2 [96] Ultrasound Breast tumor segmentation BUS CIA cGAN 2 [120] DBT and X-ray Breast mass segmentation DBT U-Net 2 [121] Cardiac image analysis…”

Section: Segmentationmentioning

confidence: 99%

Computer vision and machine learning for medical image analysis: recent advances, challenges, and way forward

Elyan¹,

Vuttipittayamongkol²,

Johnston³

et al. 2022

Art Int Surg

View full text Add to dashboard Cite

The recent development in the areas of deep learning and deep convolutional neural networks has significantly progressed and advanced the field of computer vision (CV) and image analysis and understanding. Complex tasks such as classifying and segmenting medical images and localising and recognising objects of interest have become much less challenging. This progress has the potential of accelerating research and deployment of multitudes of medical applications that utilise CV. However, in reality, there are limited practical examples being physically deployed into front-line health facilities. In this paper, we examine the current state of the art in CV as applied to the medical domain. We discuss the main challenges in CV and intelligent data-driven medical applications and suggest future directions to accelerate research, development, and deployment of CV applications in health practices. First, we critically review existing literature in the CV domain that addresses complex vision tasks, including: medical image classification; shape and object recognition from images; and medical segmentation. Second, we present an in-depth discussion of the various challenges that are considered barriers to accelerating research, development, and deployment of intelligent CV methods in real-life medical applications and hospitals. Finally, we conclude by discussing future directions.

show abstract

Section: Segmentationmentioning

confidence: 99%

Computer vision and machine learning for medical image analysis: recent advances, challenges, and way forward

Elyan¹,

Vuttipittayamongkol²,

Johnston³

et al. 2022

Art Int Surg

View full text Add to dashboard Cite

show abstract

“…en, MD-MECA module was added to the deeper U-Net network structure during the process of encoding, transmission, and decoding [24,25]. Compared with the original U-Net, the structure optimized the feature graph during the transmission of the feature graph, supervised the feature graph of the coding part in different ways, and then transmitted to the decoding part for information supplement.…”

Section: Meca and Md-mecamentioning

confidence: 99%

U-Net: A Smart Application with Multidimensional Attention Network for Remote Sensing Images

Wang

Kong

Zhang

2022

Scientific Programming

View full text Add to dashboard Cite

Building segmentation is an important step in urban planning and development. In this work, we propose a new deep learning model, namely Multidimension Attention U-Net (MDAU-Net), to accurately segment building pixels and nonbuilding pixels in remote sensing images. Furthermore, we introduce a novel Multidimension Modified Efficient Channel Attention (MD-MECA) model to enhance the network discriminative ability through considering the interdependence between feature maps. Through deepening the U-Net model to a seven-story structure, the ability to identify the building is enhanced. We apply MD-MECA to the “skip connections” in traditional U-Net, instead of simply copying the feature mapping of the contraction path to the matching extension path, to optimize the feature transfer more efficiently. The obtained results show that our proposed MDAU-Net framework achieves the most advanced performance on publicly available building data sets (i.e. the precision over the Massachusetts buildings data set and WHU data set are 97.04% and 95.68%, respectively). Furthermore, we observed that the proposed framework outperforms several state-of-the-art approaches.

show abstract

“…Lastly, most of the prior works were evaluated on datasets with small sample sizes [12], [24]- [26] or patient cohorts with specific inclusion criteria. In addition, they do not perform any external validation of their models using datasets from other hospitals.…”

Section: Related Workmentioning

confidence: 99%

3D-GMIC: an efficient deep neural network to find small objects in large 3D images

Park¹,

Chłędowski²,

Jastrzębski³

et al. 2022

Preprint

View full text Add to dashboard Cite

3D imaging enables a more accurate diagnosis by providing spatial information about organ anatomy. However, using 3D images to train AI models is computationally challenging because they consist of tens or hundreds of times more pixels than their 2D counterparts. To train with high-resolution 3D images, convolutional neural networks typically resort to downsampling them or projecting them to two dimensions. In this work, we propose an effective alternative, a novel neural network architecture that enables computationally efficient classification of 3D medical images in their full resolution. Compared to off-theshelf convolutional neural networks, 3D-GMIC uses 77.98%-90.05% less GPU memory and 91.23%-96.02% less computation. While our network is trained only with image-level labels, without segmentation labels, it explains its classification predictions by providing pixel-level saliency maps. On a dataset collected at NYU Langone Health, including 85,526 patients with full-field 2D mammography (FFDM), synthetic 2D mammography, and 3D mammography (DBT), our model, the 3D Globally-Aware Multiple Instance Classifier (3D-GMIC), achieves a breast-wise AUC of 0.831 (95% CI: 0.769-0.887) in classifying breasts with malignant findings using DBT images. As DBT and 2D mammography capture different information, averaging predictions on 2D and 3D mammography together leads to a diverse ensemble with an improved breast-wise AUC of 0.841 (95% CI: 0.768-0.895). Our model generalizes well to an external dataset from Duke University Hospital, achieving an imagewise AUC of 0.848 (95% CI: 0.798-0.896) in classifying DBT images with malignant findings.

show abstract

DBT Masses Automatic Segmentation Using U-Net Neural Networks

Cited by 18 publications

References 32 publications

Computer vision and machine learning for medical image analysis: recent advances, challenges, and way forward

Computer vision and machine learning for medical image analysis: recent advances, challenges, and way forward

U-Net: A Smart Application with Multidimensional Attention Network for Remote Sensing Images

3D-GMIC: an efficient deep neural network to find small objects in large 3D images

Contact Info

Product

Resources

About