Person Re-Identification Using Local Relation-Aware Graph Convolutional Network.

Local feature extractions have been verified to be effective for person re-identification (re-ID) in recent literature. However, existing methods usually rely on extracting local features from single part of a pedestrian while neglecting the relationship of local features among different pedestrian images. As a result, local features contain limited information from one pedestrian image, and cannot benefit from other pedestrian images. In this paper, we propose a novel approach named Local Relation-Aware Graph Convolutional Network (LRGCN) to learn the relationship of local features among different pedestrian images. In order to completely describe the relationship of local features among different pedestrian images, we propose overlap graph and similarity graph. The overlap graph formulates the edge weight as the overlap node number in the node's neighborhoods so as to learn robust local features, and the similarity graph defines the edge weight as the similarity between the nodes to learn discriminative local features. To propagate the information for different kinds of nodes effectively, we propose the Structural Graph Convolution (SGConv) operation. Different from traditional graph convolution operations where all nodes share the same parameter matrix, SGConv learns different parameter matrices for the node itself and its neighbor nodes to improve the expressive power. We conduct comprehensive experiments to verify our method on four large-scale person re-ID databases, and the overall results show LRGCN exceeds the state-of-the-art methods.

Content-Seam-Preserving Multi-Alignment Network for Visual-Sensor-Based Image Stitching.

As an important representation of scenes in virtual reality and augmented reality, image stitching aims to generate a panoramic image with a natural field-of-view by stitching multiple images together, which are captured by different visual sensors. Existing deep-learning-based methods for image stitching only conduct a single deep homography to perform image alignment, which may produce inevitable alignment distortions. To address this issue, we propose a content-seam-preserving multi-alignment network (CSPM-Net) for visual-sensor-based image stitching, which could preserve the image content consistency and avoid seam distortions simultaneously. Firstly, a content-preserving deep homography estimation was designed to pre-align the input image pairs and reduce the content inconsistency. Secondly, an edge-assisted mesh warping was conducted to further align the image pairs, where the edge information is introduced to eliminate seam artifacts. Finally, in order to predict the final stitched image accurately, a content consistency loss was designed to preserve the geometric structure of overlapping regions between image pairs, and a seam smoothness loss is proposed to eliminate the edge distortions of image boundaries. Experimental results demonstrated that the proposed image-stitching method can provide favorable stitching results for visual-sensor-based images and outperform other state-of-the-art methods.

Cross-Modality Person Re-Identification via Local Paired Graph Attention Network.

Cross-modality person re-identification (ReID) aims at searching a pedestrian image of RGB modality from infrared (IR) pedestrian images and vice versa. Recently, some approaches have constructed a graph to learn the relevance of pedestrian images of distinct modalities to narrow the gap between IR modality and RGB modality, but they omit the correlation between IR image and RGB image pairs. In this paper, we propose a novel graph model called Local Paired Graph Attention Network (LPGAT). It uses the paired local features of pedestrian images from different modalities to build the nodes of the graph. For accurate propagation of information among the nodes of the graph, we propose a contextual attention coefficient that leverages distance information to regulate the process of updating the nodes of the graph. Furthermore, we put forward Cross-Center Contrastive Learning (C3L) to constrain how far local features are from their heterogeneous centers, which is beneficial for learning the completed distance metric. We conduct experiments on the RegDB and SYSU-MM01 datasets to validate the feasibility of the proposed approach.

Admission rates of ED patients with transient ischemic attack have increased since 2000.

OBJECTIVE: A study published in December 2000 showed that 5% of patients presenting with transient ischemic attacks (TIAs) developed a stroke within 48 hours. This finding has been corroborated in several other studies. We hypothesize that, influenced by this, emergency department (ED) physicians have been more reluctant to discharge TIA patients resulting in an increase in the percentage of TIA patients admitted. METHODS: This is a retrospective cohort of consecutive ED visits. This study is conducted in 6 New Jersey EDs with annual ED visits from 25000 to 65000. Consecutive patients seen by ED physicians between January 1, 2000, and December 31, 2010, were included. We identified TIA visits using the International Classification of Diseases, Ninth Revision, code. We analyzed the admission rates for TIA testing for significant differences using the Student t test and calculated 95% confidence intervals. RESULTS: Of the 2622659 visits in the database, 8216 (0.3%) were for TIA. Females comprised 57%. There was a statistically significant increase in the annual admission rates for TIA patients from 2000 to 2010, from 70% to 91%, respectively (difference, 22%; 95% confidence interval, 18%-26% [P < .001]). Separate analysis by sex showed similar increased admission rates for females and males. CONCLUSIONS: We found that the admission rate for TIAs increased significantly from 2001 to 2010. This change in physicians' practice may be due to the body of evidence that TIA patients have a significant short-term risk of stroke.

A new divergence measure for medical image registration.

A new type of divergence measure for the registration of medical images is introduced that exploits the properties of the modified Bessel functions of the second kind. The properties of the proposed divergence coefficient are analysed and compared with those of the classic measures, including Kullback-Leibler, Renyi, and Iinfinity, divergences. To ensure its effectiveness and widespread applicability to any arbitrary set of data types, the performance of the new measure is analysed for Gaussian, exponential, and other advanced probability density functions. The results verify its robustness. Finally, the new divergence measure is used in the registration of CT to MR medical images to validate the improvement in registration accuracy.

Effects of computer-based intervention through acoustically modified speech (Fast ForWord) in severe mixed receptive-expressive language impairment: outcomes from a randomized controlled trial.

Seventy-seven children between the ages of 6 and 10 years, with severe mixed receptive-expressive specific language impairment (SLI), participated in a randomized controlled trial (RCT) of Fast ForWord (FFW; Scientific Learning Corporation, 1997, 2001). FFW is a computer-based intervention for treating SLI using acoustically enhanced speech stimuli. These stimuli are modified to exaggerate their time and intensity properties as part of an adaptive training process. All children who participated in the RCT maintained their regular speech and language therapy and school regime throughout the trial. Standardized measures of receptive and expressive language were used to assess performance at baseline and to measure outcome from treatment at 9 weeks and 6 months. Children were allocated to 1 of 3 groups. Group A (n = 23) received the FFW intervention as a home-based therapy for 6 weeks. Group B (n = 27) received commercially available computer-based activities designed to promote language as a control for computer games exposure. Group C (n = 27) received no additional study intervention. Each group made significant gains in language scores, but there was no additional effect for either computer intervention. Thus, the findings from this RCT do not support the efficacy of FFW as an intervention for children with severe mixed receptive-expressive SLI.

Efficient implementation of accurate geometric transformations for 2-D and 3-D image processing.

This paper proposes the use of a polynomial interpolator structure (based on Horner's scheme) which is efficiently realizable in hardware, for high-quality geometric transformation of two- and three-dimensional images. Polynomial-based interpolators such as cubic B-splines and optimal interpolators of shortest support are shown to be exactly implementable in the Horner structure framework. This structure suggests a hardware/software partition which can lead to efficient implementations for multidimensional interpolation.