Search

Deep learning has brought a significant progress in medical image analysis. However, their lack of interpretability might bring high risk for wrong diagnosis with limited clinical knowledge embedding. In other words, we believe it's crucial for humans to interpret how deep learning work for medical analysis, thus appropriately adding knowledge constraints to correct the bias of wrong results. With such purpose, we propose Representation Group-Disentangling Network (RGD-Net) to explain the process of feature extraction and decision making inside deep learning framework, where we completely disentangle feature space of input X-ray images into independent feature groups, and each group would contribute to diagnose of a specific disease. Specifically, we first state problem definition for interpretable prediction with auto-encoder structure. Then, group-disentangled representations are extracted from input X-ray images with the proposed Group-Disentangle Module, which constructs semantic latent space by enforcing semantic consistency of attributes. Afterwards, adversarial constricts on mapping from features to diseases are proposed to prevent model collapse during training. Finally, a novel design of local tuning medical application is proposed based on RGB-Net, which is capable to aid clinicians for reasonable diagnosis. By conducting quantity of experiments on public datasets, RGD-Net have been superior to comparative studies by leveraging potential factors contributing to different diseases. We believe our work could bring interpretability in digging inherent patterns of deep learning on medical image analysis. • RGD-Net extracts fine-grained, group-disentangled disease representations for better interpretation and prediction. • Adversarial constraint avoids shortcut problems and promotes global minimum convergence in RGD-Net. • RGD-Net significantly improves classification accuracy and disentangles information in experiments. [ABSTRACT FROM AUTHOR]

Floods result in substantial damage throughout the world every year. Accurate predictions of floods can significantly alleviate casualties and property losses. However, due to the complexity of hydrology process especially in a city with complicated pipe network, the accuracy of traditional flood forecasting models suffer from the performance degradation with the increasing of required prediction period. In the work, based on the collected historical data of Xixian City, Henan Province, China, using the Internet of Things system (IoT) in 2011-2018, a Bidirectional Gated Recurrent Unit (BiGRU) multi-step flood prediction model with attention mechanism is proposed. In our model, the attention mechanism is used to automatically adjust the matching degree between the input features and output. Besides, we use a bidirectional GRU model, which can process the input sequence from two directions of time series (chronologically and antichronologically), then merge their representations together. Compared with the prediction model using Long Short Term Memory (LSTM), our method can generate better prediction result, as can be seen from the arrival time error and peak error of floods during multi-step predictions. • A BiGRU multi-step flood prediction model based on attention mechanism was proposed. • Model tests were designed and compared with relevant models based on LSTM neural network. • The results show that the model has smaller deterministic coefficient and RMSE. • Within the allowable error range, the prediction accuracy of our model can reach 100%. [ABSTRACT FROM AUTHOR]

Inspired by instance segmentation algorithms, researchers have proposed quantity of segmentation-based methods for text detection, achieving remarkable results on scene text with arbitrary orientation and large aspect ratios. Following their success, we believe cascade architecture and extracting contextual information in multiple aspects are powerful to boost performance on the basis of segmentation-based methods, especially in decreasing false positive texts in complex natural scene. Based on such consideration, we propose a multiple-context-aware and cascade CNN structure, which appropriately encodes multiple categories of context information into a cascade R-CNN framework. Specifically, the proposed method consists of two stages, i.e., feature generation and cascade detection. During the first stage, we define ISTK (Isolated Selective Text Kernel) module to refine feature map, which sequentially encodes channel-wise and kernel-size attention information by designing multiple branches and different kernel sizes in isolate form. Afterwards, we build long-range spatial dependencies in feature map via non-local operations. Built on contextual feature map, Cascade Mask R-CNN structure progressively refines accurate boundaries of text instances with multi-stage framework. We conduct comparative experiments on ICDAR2015 and 2017-MLT datasets, where the proposed method outperform comparative methods in terms of effectiveness and efficiency measurements. [ABSTRACT FROM AUTHOR]