Search

Pathological diagnosis is the gold standard for disease assessment in clinical practice. It is conducted by inspecting the specimen at the microscopical level. Therefore, a very high-resolution pathological image that precisely describes the submicron-scale appearance is essential in the era of digital pathology, which is not easily obtained. Recently, pathological image super-resolution (SR) has shown promising prospects in bridging this gap. However, existing studies have not fully explored the peculiarity of pathological data, which contains several gradually enlarged images describing the specimen at different magnifications. In this paper, we propose a novel MMSRNet that formulates the pathological image SR in a multi-task learning way. It adds an image magnification classification branch on top of the CNN-based SR network, e.g., RCAN. Therefore, the learning objective is transformed into performing the SR while classifying the magnification as accurately as possible. The incorporated classification label guides the network to learn a more powerful feature representation. Meanwhile, the multi-task learning paradigm also encourages the joint learning of multi-scale mapping functions corresponding to multiple magnifications. It thus enables the learned model to adaptively accommodate the magnification variants, overcoming the problem that performing SR from different magnifications is treated as independent tasks in existing studies. Extensive experiments are conducted to validate the effectiveness of MMSRNet. It not only gains better performance in performing SR across magnifications and scaling factors, but also exhibits attractive plug-and-play nature when RCAN is substituted by other SR networks. The generated images are also supposed to be helpful in clinical diagnosis. • A multi-task deep learning framework for pathological image super-resolution. • Image super-resolution and magnification classification tasks are mutually reinforced. • A magnification variant multi-scale learning, well addressing the fixed magnification problem. [ABSTRACT FROM AUTHOR]

Retinal blood vessel segmentation in fundus images plays an important role in the early diagnosis and treatment of retinal diseases. In recent years, the segmentation methods based on deep neural networks have attracted the attention of experts and scholars. However, due to the complexity of the distribution of blood vessels in fundus images and the imbalance between blood vessels and background, retinal blood vessel segmentation remains challenging. In this paper, we present a retinal blood vessel segmentation method using deep neural networks. Firstly, we propose a novel deep network architecture named Bridge-net to make use of the context of the retinal blood vessels efficiently. Specifically, the architecture incorporates a recurrent neural network (RNN) into a convolutional neural network (CNN) to deliver the context and then to produce the probability map of the retinal blood vessels. Secondly, we propose a patch-based loss weight mapping by considering the distributions of different types of blood vessels to correct the imbalance, since there are large morphological differences between thick and thin blood vessels. Finally, we evaluate our method on three publicly datasets STARE, DRIVE, and CHASE _ DB1, and compare the results to eighteen state-of-the-art approaches. We also compare our method with some existing approaches on a high-resolution dataset, i.e., HRF. The results show that our method achieves better/comparable performances when compared to the existing approaches. The results on various datasets also verify the effectiveness and stability of the proposed method. • We present a novel automatic method for the segmentation of retinal blood vessels. • We propose Bridge-net by joint learning context-involved and non-context features. • We develop a patch-based loss weight mapping to correct the imbalance of the image. • We evaluate the effectiveness of the proposed method on four public datasets. • The results have verified the effectiveness and stability of the proposed method. [ABSTRACT FROM AUTHOR]

Abstract: Factors influencing the impact toughness of two impact polypropylene copolymers (IPC) with almost the same ethylene content, molecular weight and molecular weight distribution were studied by temperature gradient extraction fractionation (TGEF), scanning electron microscopy (SEM), nuclear magnetic resonance (NMR) and differential scanning calorimetry (DSC). The results indicate that poor interfacial adhesion between the disperse phase and the continuous matrix, larger dimensions and non-uniform distribution of disperse phases are main reasons for the low impact toughness of IPC B that possesses of a low content of ethylene–propylene segmented copolymer with long crystallizable PE and PP sequences as a compatibilizer between the disperse phase and the matrix. [Copyright &y& Elsevier]

• We propose an improved dataset for research on image privacy. • We build a knowledge graph that represents the relevance between the objects and image privacy. • Based on the graph, we deal with the traditional image classification problem with the novel graph-based neural network. • The introduction of the knowledge graph not only makes the model more explainable but also makes better use of the information of objects provided by the images. • Our method relies on the intrinsic relevance in the dataset rather than extra knowledge, thus is applicable for other tasks. With the development of social media and smartphones, people share their daily lives via a large number of images, but the convince also raises a problem of privacy leakage. Therefore, effective methods are needed to infer the privacy risk of images and identify images that may disclose privacy. Several works have tried to solve this problem with deep learning models. However, we know little about how the models infer the privacy label of an image, thus it is not easy to understand why the image may disclose privacy. Inspired by recent research on graph neural networks, we introduce prior knowledge to the deep models to make the inference more explainable. We propose the Graph-based neural networks for Image Privacy (GIP) to infer the privacy risk of images. The GIP mainly focuses on objects in an image, and the knowledge graph is extracted from the objects in the dataset without reliance on extra knowledge. Experimental results show that the GIP achieves higher performance compared with the object-based methods and comparable performance even compared with the multi-modal fusion method. The results show that the introduction of the knowledge graph not only makes the deep model more explainable but also makes better use of the information of objects provided by the images. Combing the knowledge graph with deep learning is a promising way to help protect image privacy that is worth exploring. [ABSTRACT FROM AUTHOR]

Abstract Lung cancer is one of the leading causes of cancer-related death worldwide. Early diagnosis can effectively reduce the mortality, and computer-aided diagnosis (CAD) as an important way to assist doctors has developed rapidly. In particular, automated pulmonary nodule detection in computed tomography (CT) images is crucial to CAD. It is a challenging task to quickly locate the exact positions of lung nodules. In this paper, a novel automated pulmonary nodule detection framework with 2D convolutional neural network (CNN) is proposed to assist the CT reading process. Firstly, we adjust the structure of Faster R-CNN with two region proposal networks and a deconvolutional layer to detect nodule candidates, and then three models are trained for three kinds of slices for later result fusion. Secondly, a boosting architecture based on 2D CNN is designed for false positive reduction, which is a classifier to distinguish true nodules from the candidates. The misclassified samples are still kept for retraining a model which boosts the sensitivity for pulmonary nodule detection. Finally, the results of these networks are fused to vote out the final classification results. Extensive experiments are conducted on LUNA16, and the sensitivity of nodule candidate detection achieves 86.42%. For the false positive reduction, the sensitivity reaches 73.4% and 74.4% at 1/8 and 1/4 FPs/scan, respectively. It illustrates that the proposed method can obviously achieve accurate pulmonary nodule detection. [ABSTRACT FROM AUTHOR]

Abstract Distant supervised relation extraction is an important task in the field of natural language processing. There are two main shortcomings for most state-of-the-art methods. One is that they take all sentences of an entity pair as input, which would result in a large computational cost. But in fact, few of most relevant sentences are enough to recognize the relation of an entity pair. To tackle these problems, we propose a novel hierarchical selective attention network for relation extraction under distant supervision. Our model first selects most relevant sentences by taking coarse sentence-level attention on all sentences of an entity pair and then employs word-level attention to construct sentence representations and fine sentence-level attention to aggregate these sentence representations. Experimental results on a widely used dataset demonstrate that our method performs significantly better than most of existing methods. [ABSTRACT FROM AUTHOR]

Accurate skin lesion segmentation in dermoscopic images is crucial to the early diagnosis of skin cancers. However, it remains a challenging task due to fuzzy lesion boundaries, irregular lesion shapes, and the existence of various interference factors. In this paper, a novel Attention Synergy Network (AS-Net) is developed to enhance the discriminative ability for skin lesion segmentation by combining both spatial and channel attention mechanisms. The spatial attention path captures lesion-related features in the spatial dimension while the channel attention path selectively emphasizes discriminative features in the channel dimension. The synergy module is designed to optimally integrate both spatial and channel information, and a weighted binary cross-entropy loss function is introduced to emphasize the foreground lesion region. Comprehensive experiments indicate that our proposed model achieves the state-of-the-art performance with the highest overall score in the ISIC2017 challenge, and outperforms several popular deep neural networks on both ISIC2018 and PH2 datasets. • Proposes AS-Net to enhance the discriminative ability for skin lesion segmentation. • Combines spatial and channel attentions for feature learning of skin lesions. • A weighted loss function is presented to emphasize the skin lesion region. • Our model outperforms state-of-the-art methods on three publicly available datasets. [ABSTRACT FROM AUTHOR]

Highlights: [•] PM is an ideal candidate for co-digestion with MAW. [•] The optimal COD ratio of PM:MAW for anaerobic digestion is 1.0:1.5. [•] The structure and composition of bacterial communities varied in the early and late stage. [•] The type of main OTUs did not differ much among tanks with different ratios of PM to MAW, only the proportion is different. [ABSTRACT FROM AUTHOR]

• We propose an efficient 3D residual neural network for brain tumor segmentation. • We propose a fusion loss function based on Dice and Cross-entropy. • We introduce a concise but effective post-processing method. • The evaluation is performed on the BRATS 2018 dataset. • The results demonstrate that our method outperforms the state-of-the-art approaches. Brain tumors are the most aggressive and mortal cancers, which lead to short life expectancy. A reliable and efficient automatic or semi-automatic segmentation method is significant for clinical practice. In recent years, deep learning-based methods achieve great success in brain tumor segmentation. However, due to the limitation of parameters and computational complexity, there is still much room for improvement in these methods. In this paper, we propose an efficient 3D residual neural network (ERV-Net) for brain tumor segmentation, which has less computational complexity and GPU memory consumption. In ERV-Net, a computation-efficient network, 3D ShuffleNetV2, is firstly utilized as encoder to reduce GPU memory and improve the efficiency of ERV-Net, and then the decoder with residual blocks (Res-decoder) is introduced to avoid degradation. Furthermore, a fusion loss function, which is composed of Dice loss and Cross-entropy loss, is developed to solve the problems of network convergence and data imbalance. Moreover, a concise and effective post-processing method is proposed to refine the coarse segmentation result of ERV-Net. The experimental results on the dataset of multimodal brain tumor segmentation challenge 2018 (BRATS 2018) demonstrate that ERV-Net achieves the best performance with Dice of 81.8%, 91.21% and 86.62% and Hausdorff distance of 2.70 mm, 3.88 mm and 6.79 mm for enhancing tumor, whole tumor and tumor core, respectively. Besides, ERV-Net also achieves high efficiency compared to the state-of-the-art methods. [ABSTRACT FROM AUTHOR]