Medical image quality is highly relative to clinical diagnosis and treatment, leading to a popular research topic of medical image denoising. Image denoising based on deep learning methods has attracted considerable attention owing to its excellent ability of automatic feature extraction. Most existing methods for medical image denoising adapted to certain types of noise have difficulties in handling spatially varying noise; meanwhile, image detail losses and structure changes occurred in the denoised image. Considering image context perception and structure preserving, this paper firstly introduces a medical image denoising method based on conditional generative adversarial network (CGAN) for various unknown noises. In the proposed architecture, noise image with the corresponding gradient image is merged as network conditional information, which enhances the contrast between the original signal and noise according to the structural specificity. A novel generator with residual dense blocks makes full use of the relationship among convolutional layers to explore image context. Furthermore, the reconstruction loss and WGAN loss are combined as the objective loss function to ensure the consistency of denoised image and real image. A series of experiments for medical image denoising are conducted with the denoising results of and on JSRT datasets and and on LIDC datasets. Compared with the state-of-the-art methods, the superior performance of the proposed method is outstanding.

The content-based video retrieval is a popular topic in computer vision filed, especially action retrieval. This work proposes a novel and effective video representation module for content-based action retrieval framework, in which feature learning can be conducted with complementary information and intrinsic structure that the relationship between appearance feature and geometry can be preserved. Based on multi-view analysis and graph embedding, the target features are generated to minimize the inter-class discrepancy and maximize intra-class discrimination. Applied to the content-based retrieval task, the proposed method can be combined with Euclidean distance for the comparison of low-dimensional features. Demonstrated in the extensive experiments on the benchmark datasets, the performance of the proposed framework is superior to the state-of-the-art methods.

To solve the problem that the traditional convolutional neural network uses the pooling layers to reduce the image feature dimensions, which leads to information loss and affects the accuracy of right ventricular segmentation, a right ventricular segmentation method based on U-Net improved network is proposed. The dense blocks are used to combine the bottom features of the contracting path, and shortcut connections are used to connect the low-level features and high-level features on the expanding path, which increase the reusability of features. Depthwise separable weighted convolutions are used to enhance the edge detail information and improve the possibility of information reconstruction. An improved shinkage loss function is proposed to solve the problem of unbalanced positive and negative samples. Finally, RVSC-MACCAI 2012 datasets are used in the comparison experiments of different models, and the results show the effectiveness of the improved algorithm with the Dice coefficient of 0.90 and Hausdorff distance of 6.42.

In clinical applications, the task of medical image quality assessment is done by radiologists and the existence of subjective factors leads to unstable the accuracy of the assessment. Deep learning methods bring the opportunity for objective assessment image quality. In order to avoid the overfitting problem of deep learning methods in small-sample data assessment, this paper proposes a medical image quality assessment method based on residual learning. This method adds a residual module to the traditional quality assessment network, so that the problem of overfitting can be effectively alleviated in the process of deepening the network structure. At the same time, this paper initially establishes a chest X-ray image quality assessment dataset. In the experiment, the correlation coefficients SRCC and PLCC of this method reaches 0.9692 and 0.9589, respectively, and the mean square error is 0.6697. Compared with the six assessment methods, three indicators are optimal. The method proposed in this paper can meet the clinical assessment needs of medical image quality, and the effectiveness of the method has been fully verified.

Multimodal medical images are useful for observing tissue structure clearly in clinical practice. To integrate multimodal information, multimodal registration is significant. The entropy-based registration applies a structure descriptor set to replace the original multimodal image and compute similarity to express the correlation of images. The accuracy and converging rate of the registration depend on this set. We propose a new method, logarithmic fuzzy entropy function, to compute the descriptor set. It is obvious that the proposed method can increase the upper bound value from log(r) to log(r) + ∆(r) so that a more representative structural descriptor set is formed. The experiment results show that our method has faster converging rate and wider quantified range in multimodal medical images registration.

COVID-19 has become the largest public health event worldwide since its outbreak, and early detection is a prerequisite for effective treatment. Chest X-ray images have become an important basis for screening and monitoring the disease, and deep learning has shown great potential for this task. Many studies have proposed deep learning methods for automated diagnosis of COVID-19. Although these methods have achieved excellent performance in terms of detection, most have been evaluated using limited datasets and typically use a single deep learning network to extract features. To this end, the dual asymmetric feature learning network (DAFLNet) is proposed, which is divided into two modules, DAFFM and WDFM. DAFFM mainly comprises the backbone networks EfficientNetV2 and DenseNet for feature fusion. WDFM is mainly for weighted decision-level fusion and features a new pretrained network selection algorithm (PNSA) for determination of the optimal weights. Experiments on a large dataset were conducted using two schemes, DAFLNet-1 and DAFLNet-2, and both schemes outperformed eight state-of-the-art classification techniques in terms of classification performance. DAFLNet-1 achieved an average accuracy of up to 98.56 for the triple classification of COVID-19, pneumonia, and healthy images.