Automatic segmentation of left ventricle in cardiac cine MRI images based on deep learning

Zhou, Tian; Icke, Ilknur; Dogdas, Belma; Parimal, Sarayu; Sampath, Smita; Forbes, Joseph; Bagchi, Ann D.; Chin, Chih‐Liang; Chen, Antong

doi:10.1117/12.2253963

Cited by 7 publications

(6 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to evaluate generalization, we use an experimental setup that considers several different domains, each coming from different video anomaly detection datasets. Figure 2 depicts this setup: all datasets are individually mapped to a feature space using the same pre-trained VGG-19 [34] extraction is the generalization of the feature space produced for data not seen within the same visual domain [36,37]. Among current deep learning techniques, CNNs are widely used to compute feature space representations by sharing information and internal connections [38].…”

Section: Methodsmentioning

confidence: 99%

“…To lessen this hindrance, data-driven approaches to feature extraction can be employed, of which deep learning was particularly shown to produce suitable representations. One of the main advantages presented by methods of feature learning in relation to handcrafted extraction is the generalization of the feature space produced for data not seen within the same visual domain [36,37]. Among current deep learning techniques, CNNs are widely used to compute feature space representations by sharing information and internal connections [38].…”

Section: Feature Extractionmentioning

confidence: 99%

See 1 more Smart Citation

Generalization of feature embeddings transferred from different video anomaly detection domains

Santos

Ribeiro

Ponti

2019

Journal of Visual Communication and Image Representation

View full text Add to dashboard Cite

Detecting anomalous activity in video surveillance often involves using only normal activity data in order to learn an accurate detector. Due to lack of annotated data for some specific target domain, one could employ existing data from a source domain to produce better predictions. Hence, transfer learning presents itself as an important tool. But how to analyze the resulting data space? This paper investigates video anomaly detection, in particular feature embeddings of pre-trained CNN that can be used with non-fully supervised data. By proposing novel cross-domain generalization measures, we study how source features can generalize for different target video domains, as well as analyze unsupervised transfer learning. The proposed generalization measures are not only a theorical approach, but show to be useful in practice as a way to understand which datasets can be used or transferred to describe video frames, which it is possible to better discriminate between normal and anomalous activity.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Feature Extractionmentioning

confidence: 99%

Generalization of feature embeddings transferred from different video anomaly detection domains

Santos

Ribeiro

Ponti

2019

Journal of Visual Communication and Image Representation

View full text Add to dashboard Cite

show abstract

“…The approach is divided into two categories: LV localization and contour segmentation, including shape feature detection [33], [34], LV segmentation and function estimation using deep learning [35]- [39]; automatic identification of blood vessel using a cascading classifier [40], diffusion-based unsupervised clustering technique for Myocardial motion patterns classification [41]; CNN and U-Net approach [42], Multi-input fusion network [43], cardiac motion measurement by used algorithm namely surface structure feature matching [44], deep earning method used deformable, level set and threshold method for automatic LV contour segmentation [45]- [48]. Another major challenge is a region of interest (ROI) for automatic contour segmentation.…”

Section: Fully Automatic Approachmentioning

confidence: 99%

Segmentation Model Approaches using Cardiac Magnetic Resonance Images: A Review

Khan¹

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Both training and testing images are preprocessed following the steps described in [12], where each slice is resampled to 1×1 mm 2 pixel size and cropped to a 184×184 matrix from the center. Further LV localization is performed by motion analysis and Hough transform for detecting circles, such that the bounding box enclosing the LV area can be defined with a set margin.…”

Section: Image Preprocessingmentioning

confidence: 99%

“…Inspired by these RNN-based approaches, especially the work in [11], we propose a multi-level convolutional LSTM (ConvLSTM) approach for the automatic segmentation of LV myocardium. To develop the ConvLSTM model, a ResNet-56 CNN model [12] is trained first, and the LVrelated image features at the low-and high-resolution levels are extracted separately, each for training one LSTM model. Using a leave-one-out approach, we compare the proposed model with a one-level ConvLSTM and a CNN model by evaluating them in a dataset of which the LV myocardium is manually delineated.…”

Section: Introductionmentioning

confidence: 99%

A multi-level convolutional LSTM model for the segmentation of left ventricle myocardium in infarcted porcine cine MR images

Zhang

Icke

Dogdas

et al. 2018

2018 IEEE 15th International Symposium on Biomedical Imaging (ISBI 2018)

Self Cite

View full text Add to dashboard Cite

Automatic segmentation of left ventricle (LV) myocardium in cardiac short-axis cine MR images acquired on subjects with myocardial infarction is a challenging task, mainly because of the various types of image inhomogeneity caused by the infarctions. Among the approaches proposed to automate the LV myocardium segmentation task, methods based upon deep convolutional neural networks (CNN) have demonstrated their exceptional accuracy and robustness in recent years. However, most of the CNN-based approaches treat the frames in a cardiac cycle independently, which fails to capture the valuable dynamics of heart motion. Herein, an approach based on recurrent neural network (RNN), specifically a multi-level convolutional long short-term memory (ConvLSTM) model, is proposed to take the motion of the heart into consideration. Based on a ResNet-56 CNN, LV-related image features in consecutive frames of a cardiac cycle are extracted at both the low-and highresolution levels, which are processed by the corresponding multi-level ConvLSTM models to generate the myocardium segmentations. A leave-one-out experiment was carried out on a set of 3,600 cardiac cine MR slices collected in-house for 8 porcine subjects with surgically induced myocardial infarction. Compared with a solely CNN-based approach, the proposed approach demonstrated its superior robustness against image inhomogeneity by incorporating information from adjacent frames. It also outperformed a one-level ConvLSTM approach thanks to its capabilities to take advantage of image features at multiple resolution levels.Index Terms-Cardiac cine MRI, convolutional neural network, recurrent neural network, LSTM

show abstract

Automatic segmentation of left ventricle in cardiac cine MRI images based on deep learning

Cited by 7 publications

References 3 publications

Generalization of feature embeddings transferred from different video anomaly detection domains

Generalization of feature embeddings transferred from different video anomaly detection domains

Segmentation Model Approaches using Cardiac Magnetic Resonance Images: A Review

A multi-level convolutional LSTM model for the segmentation of left ventricle myocardium in infarcted porcine cine MR images

Contact Info

Product

Resources

About