A convolutional neural network‐based model observer for breast CT images

Kim, Gihun; Han, Minah; Shim, Hyunjung; Baek, Jongduk

doi:10.1002/mp.14072

Cited by 23 publications

(38 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…CNNs also have strong representational power to learn non-linear functions, which makes it possible to develop a single model that generalizes to different imaging conditions. It has been shown that CNNs can approximate an ideal model observer [12], [13] and generally perform better than human observers [14] when they are trained with ground truth label (e.g., 1 for all signal-present images and 0 for all signal-absent images). Therefore, to implement CNNbased anthropomorphic model observers, it is necessary to reflect the human inefficiency in CNNs.…”

Section: Introductionmentioning

confidence: 99%

A Convolutional Neural Network-Based Anthropomorphic Model Observer for Signal Detection in Breast CT Images Without Human-Labeled Data

2020

Self Cite

View full text Add to dashboard Cite

Various imaging parameters in X-ray computed tomography (CT) should be examined and optimized by task-based assessment of human observer performance. Recently, convolutional neural networks (CNNs) have been introduced as anthropomorphic model observers. However, when humanlabeled data are not available or limited, CNNs with an existing training strategy do not produce good performance agreement with human observers. The purpose of this study is to propose new training strategies for a CNN-based anthropomorphic model observer without human-labeled data for signal-knownexactly and background-known-statistically detection tasks. We acquired cone-beam CT projection data of breast background volume and reconstructed the projection data using the Feldkamp-Davis-Kress algorithm with 12 different imaging conditions including viewing image planes. Training data for the CNN were labeled utilizing conventional model observers. We employed an early stopping rule to reflect internal noise during the CNN training. To examine the CNN performance, we used three different training-testing schemes. The performance agreement between the human and model observers was measured via a Bland-Altman plot, the root-mean-squared error (RMSE), and the Pearson's correlation coefficient (r) of their proportion correct values. Throughout the three different training-testing schemes, CNNs with the proposed training strategies yielded narrower limits of agreements (with a bias lower than 0.03) and higher scores in both RMSE and r than the conventional model observers. This indicates the proposed training strategies enable the CNN-based anthropomorphic model observer to have good performance agreement with human observers and generalize better to different imaging conditions than conventional anthropomorphic model observers. INDEX TERMS Anthropomorphic model observer, breast cone-beam CT, convolutional neural network, data labeling, early stopping, signal detection.

show abstract

Section: Introductionmentioning

confidence: 99%

A Convolutional Neural Network-Based Anthropomorphic Model Observer for Signal Detection in Breast CT Images Without Human-Labeled Data

2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…Data augmentation was used to increase both the size and the diversity of the training data set. Gaussian filters ( 16 ) are used to transform the input image in the data augmentation step. After data augmentation, all images were resized to 512 × 512 pixels, and a mean normalization was performed as follows:…”

Section: Methodsmentioning

confidence: 99%

Convolutional Neural Network-Based Computer-Assisted Diagnosis of Hashimoto’s Thyroiditis on Ultrasound

Zhao

Kang

Qian

et al. 2021

The Journal of Clinical Endocrinology &Amp; Metabolism

View full text Add to dashboard Cite

Purpose This study investigates the efficiency of deep learning models in the automated diagnosis of Hashimoto’s thyroiditis (HT) using real-world ultrasound data from ultrasound examinations by computer-assisted diagnosis (CAD) with artificial intelligence. Methods We retrospectively collected ultrasound images from patients with and without HT from two hospitals in China between September 2008 and February 2018. Images were divided into a training set (80%) and a validation set (20%). We ensembled nine convolutional neural networks (CNNs) as the final model (CAD-HT) for HT classification. The model’s diagnostic performance was validated and compared from two hospital validation sets. We also compared the accuracy of CAD-HT against seniors/junior radiologists. Subgroup analysis of CAD-HT performance in different thyroid hormone levels, such as hyperthyroidism, hypothyroidism, and euthyroidism, was also evaluated. Results 39280 ultrasound images from 21118 patients were included in this study. The accuracy, sensitivity, specificity of the ensemble HT-CAD model were 0.892, 0.890 and 0.895, respectively. HT-CAD performance between two hospitals was not significantly different. The HT-CAD model achieved a higher performance (P < 0.001) when compared to senior radiologists, with a nearly 9% accuracy improvement. HT-CAD had almost similar accuracy (from 0.871 to 0.894) among the three differences of thyroid hormone level subgroups. Conclusion The HT-CAD strategy based on CNN significantly improved the radiologists’ diagnostic accuracy on HT. Our model demonstrates good performance and robustness in different hospitals and for different thyroid hormone levels.

show abstract

“…Furthermore, because we conducted a binary classification task, the network architecture was much more cost‐effective than a VGG network, as it reduced the number of network parameters and training images required. In our previous study, 21 we found that a CNN‐based model observer provided better detection performance than a conventional linear model observer when it was trained using more than 10 000 image pairs. Thus, we implemented the observer loss using Mayo Clinic data, which contained more than 10 000 pairs of image patches.…”

Section: Introductionmentioning

confidence: 93%

“…For all the network architectures in Figure 3, there were no improvements in the detection accuracies when the number of image pairs used for observer network training was more than 10 000 in the SKE detection task, which was consistent with our previous findings. 21 For SKS detection, there were no improvements in the detection accuracies when the number of image pairs used for training was more than 20 000 because this task was more complex than SKE detection. The weights of the convolution layers were initialized to random values sampled from a Gaussian distribution with zero mean and a standard deviation of √ 2∕N, where N is the number of incoming nodes (i.e., He normalization).…”

Section: Training For the Observer Networkmentioning

confidence: 99%

Low‐dose CT denoising via convolutional neural network with an observer loss function

2021

Self Cite

View full text Add to dashboard Cite

Convolutional neural network (CNN)-based denoising is an effective method for reducing complex computed tomography (CT) noise. However, the image blur induced by denoising processes is a major concern.The main source of image blur is the pixel-level loss (e.g., mean squared error [MSE] and mean absolute error [MAE]) used to train a CNN denoiser. To reduce the image blur, feature-level loss is utilized to train a CNN denoiser. A CNN denoiser trained using visual geometry group (VGG) loss can preserve the small structures, edges, and texture of the image.However, VGG loss, derived from an ImageNetpretrained image classifier, is not optimal for training a CNN denoiser for CT images. ImageNet contains natural RGB images, so the features extracted by the ImageNet-pretrained model cannot represent the characteristics of CT images that are highly correlated with diagnosis. Furthermore, a CNN denoiser trained with VGG loss causes bias in CT number. Therefore, we propose to use a binary classification network trained using CT images as a feature extractor and newly define the feature-level loss as observer loss. Methods: As obtaining labeled CT images for training classification network is difficult, we create labels by inserting simulated lesions. We conduct two separate classification tasks, signal-known-exactly (SKE) and signal-knownstatistically (SKS), and define the corresponding feature-level losses as SKE loss and SKS loss, respectively. We use SKE loss and SKS loss to train CNN denoiser. Results: Compared to pixel-level losses, a CNN denoiser trained using observer loss (i.e., SKE loss and SKS loss) is effective in preserving structure, edge, and texture. Observer loss also resolves the bias in CT number, which is a problem of VGG loss. Comparing observer losses using SKE and SKS tasks, SKS yields images having a more similar noise structure to reference images. Conclusions: Using observer loss for training CNN denoiser is effective to preserve structure, edge, and texture in denoised images and prevent the CT number bias. In particular, when using SKS loss, denoised images having a similar noise structure to reference images are generated.

show abstract

A convolutional neural network‐based model observer for breast CT images

Cited by 23 publications

References 44 publications

A Convolutional Neural Network-Based Anthropomorphic Model Observer for Signal Detection in Breast CT Images Without Human-Labeled Data

A Convolutional Neural Network-Based Anthropomorphic Model Observer for Signal Detection in Breast CT Images Without Human-Labeled Data

Convolutional Neural Network-Based Computer-Assisted Diagnosis of Hashimoto’s Thyroiditis on Ultrasound

Low‐dose CT denoising via convolutional neural network with an observer loss function

Contact Info

Product

Resources

About