A novel semantic loop closure detection method (SLCD) is proposed in this paper for visual simultaneous localization and mapping (V-SLAM) systems. SLCD aims to relieve the instance-level semantic inconsistency issue that arose from dynamic industrial scenes (e.g., autonomous driving in big cities). As the first step in this direction, SLCD fully exploits both low and high-level video frame information, in a coarse-to-fine way. In SLCD, we adopt a convolutional neural network(CNN)-based object detection to acquire object information from the consecutive frames. Meanwhile, we perform a bag of visual words (BoVW)-based similarity calculation to narrow the frames to coarse loop closure candidates. For these candidates, we perform an object matching on them to find their semantic inconsistency cases and remove involved semantic inconsistencies according to their cases. Then we recalculate the similarity scores for these candidates. Finally, loop closures are determined by the similarity scores and a geometrical verification. Favorable performance of the proposed method is demonstrated by comparing it to other state-of-the-art methods using data from several public datasets and our new Dynamic Scenes dataset.

A proper-orthogonal-decomposition analysis was performed on the turbulent flow in the near-wake region of an S805 airfoil in deep stall at an angle of attack of 30 deg. The flow was measured using tomographic particle image velocimetry at a Reynolds number of 4600. Instantaneous turbulence structures, which were significant contributors to the first two proper-orthogonal-decomposition modes, were studied. These structures included the large-scale Kármán vortex and small-scale shear-layer vortices that were interacting with the Kármán vortex. An interesting correspondence was observed between the rotational vectors in proper-orthogonal-decomposition modes 4 and 5 and the locations of the leading-edge shear-layer vortices in the flowfields, which were major contributors to these two modes. Similarity was found in the autospectral functions computed from the velocity fields, which were significant contributors to the first five proper-orthogonal-decomposition modes. The wavelengths corresponding to the peak values of autospectral functions could be related to the size of the induced flows by the Kármán vortex and the streamwise spacing of the shear-layer vortices.

An image can be described in terms of appearance frequency of visual words. This representation is implemented in bag-of-visual-words (BoVW)-based loop closure detection for its efficiency and effectiveness. However, traditional BoVW-based approaches are strongly affected by false positive loops due to scene ambiguity caused by redundant words in the vocabulary and fail to detect bidirectional loops in monocular mode. Aiming at overcoming these problems, we propose a novel vocabulary construction algorithm named hierarchical sequential information bottleneck (HsIB) by leveraging the maximization of mutual information (MMI) mechanism. First, feature descriptors are extracted from training images for visual vocabulary construction. Second, HsIB extracts discriminative yet informative visual words through the MMI mechanism in vocabulary construction, which treats feature descriptors clustering as a process of data compression. Finally, the clustering process reaches a tradeoff between compactness and discrimination and improves the performance of traditional BoVW-based loop closure detection. The proposed method is compared with state-of-the-art methods on publicly available datasets. We also create a challenging dataset to further evaluate the performance of HsIB on bidirectional loops. To the best of our knowledge, we are the first to implement information bottleneck (IB) method in visual-SLAM (vSLAM) loop closure detection, and we obtain impressive results.

The deep convolutional neural network (ConvNet) achieves significant segmentation performance on medical images of various modalities. However, the isolated errors in a large testing set with various tumor conditions are not acceptable in clinical practice. This is usually caused in inadequate training and noise inherent during data collection, which are recognized as epistemic and aleatoric uncertainties in deep learning-based approaches. In this paper, we analyze the two types of uncertainties in medical image segmentation tasks and propose a reduction method by training models with data augmentation. The shelter zones in images are reduced with 2D imaging on surfaces of different angles from 3D organs. Rotation transformation and noise are estimated by Monte Carlo simulation with prior parameter distributions, and the aleatoric uncertainty is quantized in this process. Experiments on segmentation of computed tomography images demonstrate that overconfident incorrect predictions are reduced through uncertainty reduction and that our method outperforms prediction baselines based on epistemic and aleatoric estimation.

Deep learning methods have been migrated to rectal cancer staging as a classification process based on magnetic resonance images (MRIs). Typical approaches suffer from the imperceptible variation of images from different stage. The data augmentation also introduces scale invariance and rotation consistency problems after converting MRIs to 2D visible images. Moreover, the correctly labeled images are inadequate since T-staging requires pathological examination for confirmation. It is difficult for classification model to characterize the distinguishable features with limited labeled data. In this article, Laplace of Gaussian (LoG) filter is used to enhance the texture details of converted MRIs and we propose a new method named LoG-staging to predict the T stages of rectal cancer patients. We first use the LoG operator to clarify the fuzzy boundaries of rectal cancer cell proliferation. Then, we propose a new feature clustering method by leveraging the maximization of mutual information (MMI) mechanism which jointly learns the parameters of a neural network and the cluster assignments of features. The assignments are used as labels for the next round of training, which compensate the inadequacy of labeled training data. Finally, we experimentally verify that the LoG-staging is more accurate than the nonlinear dimensionality reduction in predicting the T stages of rectal cancer. We innovatively implement information bottleneck (IB) method in T-staging of rectal cancer based on image classification and impressive results are obtained.