Data imbalance issue generally exists in most medical image analysis problems, and may be getting important with the popularization of data-hungry deep learning paradigms. We explore the cutting-edge wasserstein generative adversarial networks (WGAN) to address the data imbalance problem with over-sampling on the minority classes. The WGAN can estimate the underlying distribution of a minority class to synthesize more plausible and helpful samples for the classification model. In this paper, the WGAN-based over-sampling technique is applied to augment the data to balance for the fine-grained classification of 7 semantic attributes of lung nodules in CT images. The fine-grained classification is carried out with a normal convolutional neural network (CNN). To further illustrate the efficacy of the WGAN-based over-sampling technique, the conventional data augmentation method commonly used in many deep learning works, the generative adversarial networks (GAN) and the deep convolutional generative adversarial networks (DCGAN) are implemented for comparison. The whole schemes of the minority over-sampling and fine-grained classification are tested with the public Lung Imaging Database Consortium (LIDC) dataset. The experimental results suggest that the WGAN-based over-sampling technique can synthesize helpful samples for the minority classes to assist the training of the CNN model and to boost the fine-grained classification performance better than the conventional data augmentation method and the two schemes of the GAN and DCGAN techniques do. It may thus suggest that the WGAN technique offers an alternative methodological option for the further deep learning on imbalanced classification studies.
15 December 2020 in BMC Medical Informatics and Decision Making
Background: Pneumothorax (PTX) may cause a life-threatening medical emergency with cardio-respiratory collapse that requires immediate intervention and rapid treatment. The screening and diagnosis of pneumothorax usually rely on chest radiographs. However, the pneumothoraces in chest X-rays may be very subtle with highly variable in shape and overlapped with the ribs or clavicles, which are often difficult to identify. Our objective was to create a large chest X-ray dataset for pneumothorax with pixel-level annotation and to train an automatic segmentation and diagnosis framework to assist radiologists to identify pneumothorax accurately and timely. Methods: In this study, an end-to-end deep learning framework is proposed for the segmentation and diagnosis of pneumothorax on chest X-rays, which incorporates a fully convolutional DenseNet (FC-DenseNet) with multi-scale module and spatial and channel squeezes and excitation (scSE) modules. To further improve the precision of boundary segmentation, we propose a spatial weighted cross-entropy loss function to penalize the target, background and contour pixels with different weights. Results: This retrospective study are conducted on a total of eligible 11,051 front-view chest X-ray images (5566 cases of PTX and 5485 cases of Non-PTX). The experimental results show that the proposed algorithm outperforms the five state-of-the-art segmentation algorithms in terms of mean pixel-wise accuracy (MPA) with $$0.93\pm 0.13$$ and dice similarity coefficient (DSC) with $$0.92\pm 0.14$$ , and achieves competitive performance on diagnostic accuracy with 93.45% and $$F_1$$ -score with 92.97%. Conclusion: This framework provides substantial improvements for the automatic segmentation and diagnosis of pneumothorax and is expected to become a clinical application tool to help radiologists to identify pneumothorax on chest X-rays.
Traditional super-resolution reconstruction methods for flow fields use end-to-end mapping to determine the relationship between high- and low-resolution flow field data. The reconstruction quality of these methods depends on the accuracy of the low-resolution data. Ensuring the accuracy of low-resolution data has, thus, become a precondition for super-resolution tasks, and it imposes strict limitations on the applicability of super-resolution reconstruction methods in practical engineering applications. This paper proposes a flow field super-resolution reconstruction method coupled with feature recognition (FRNet) to reduce the dependence on the accuracy of low-resolution data. FRNet uses a feature extractor with identification capabilities to determine the effectiveness of low-resolution flow field characteristics. It recognizes the effective characteristics using a feature distance distribution. Meanwhile, a representation of the obstacle shape and freestream information is introduced to compensate for invalid features and to suppress the influence of low-precision flow field characteristics on the reconstruction results. Different downsampling factors, different density grids, and noise are used to simulate a variety of engineering application scenarios to verify the effectiveness and applicability of the proposed method. The results demonstrate that FRNet has significant advantages over traditional super-resolution reconstruction methods. Our method does not rely on the accuracy of low-resolution data and can effectively mitigate the impact of low-resolution flow field data that do not conform to physical phenomena. This characteristic allows FRNet to exhibit outstanding performance when handling flow field data affected by noise from wind tunnel wall and rack interferences. Consequently, FRNet should prove highly beneficial for the optimization of complex flow fields using super-resolution reconstruction methods.
With the development of deep learning and artificial intelligence, deep neural networks are increasingly being applied for natural language processing tasks. However, the majority of research on natural language processing focuses on alphabetic languages. Few studies have paid attention to the characteristics of ideographic languages, such as the Chinese language. In addition, the existing Chinese processing algorithms typically regard Chinese words or Chinese characters as the basic units while ignoring the information contained within the deeper architecture of Chinese characters. In the Chinese language, each Chinese character can be split into several components, or strokes. This means that strokes are the basic units of a Chinese character, in a manner similar to the letters of an English word. Inspired by the success of character-level neural networks, we delve deeper into Chinese writing at the stroke level for Chinese language processing. We extract the basic features of strokes by considering similar Chinese characters to learn a continuous representation of Chinese characters. Furthermore, word embeddings trained at different granularities are not exactly the same. In this paper, we propose an algorithm for combining different representations of Chinese words within a single neural network to obtain a better word representation. We develop a Chinese word representation service for several natural language processing tasks, and cloud computing is introduced to deal with preprocessing challenges and the training of basic representations from different dimensions.