Abstract. Census transform (CT), a stereo matching algorithm, has a strong advantage in radial distortion and brightness changes. However, CT is noise-sensitive because it compares the brightness of a single central pixel based on the brightness values of neighborhood pixels within a matching window. Star-census transform, which compares the brightness of pixels separated by a certain distance along a symmetrical pattern within the matching window, is presented. The proposed method can select the distance between the pixels for comparison and comparison patterns. The experiment results show that the proposed method yields a better performance than the previous CT methods. © The Authors. Published by SPIE under a Creative Commons Attribution 3.0 Unported License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI.

Deep learning has been utilized in end-to-end camera pose estimation. To improve the performance, we introduce a camera pose estimation method based on a 2D-3D matching scheme with two convolutional neural networks (CNNs). The scene is divided into voxels, whose size and number are computed according to the scene volume and the number of 3D points. We extract inlier points from the 3D point set in a voxel using random sample consensus (RANSAC)-based plane fitting to obtain a set of interest points consisting of a major plane. These points are subsequently reprojected onto the image using the ground truth camera pose, following which a polygonal region is identified in each voxel using the convex hull. We designed a training dataset for 2D–3D matching, consisting of inlier 3D points, correspondence across image pairs, and the voxel regions in the image. We trained the hierarchical learning structure with two CNNs on the dataset architecture to detect the voxel regions and obtain the location/description of the interest points. Following successful 2D–3D matching, the camera pose was estimated using n-point pose solver in RANSAC. The experiment results show that our method can estimate the camera pose more precisely than previous end-to-end estimators.

Abstract. A method to estimate the environmental illumination distribution of a scene with gradient-based ray and candidate shadow maps is presented. In the shadow segmentation stage, we apply a Canny edge detector to the shadowed image by using a three-dimensional (3-D) augmented reality (AR) marker of a known size and shape. Then the hierarchical tree of the connected edge components representing the topological relation is constructed, and the connected components are merged, taking their hierarchical structures into consideration. A gradient-based ray that is perpendicular to the gradient of the edge pixel in the shadow image can be used to extract the shadow regions. In the light source detection stage, shadow regions with both a 3-D AR marker and the light sources are partitioned into candidate shadow maps. A simple logic operation between each candidate shadow map and the segmented shadow is used to efficiently compute the area ratio between them. The proposed method successively extracts the main light sources according to their relative contributions on the segmented shadows. The proposed method can reduce unwanted effects due to the sampling positions in the shadow region and the threshold values in the shadow edge detection. © The Authors. Published by SPIE under a Creative Commons Attribution 3.0 Unported License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI.

We introduce a deep-learning neural network model that uses electrocardiogram (ECG) data to predict coronary artery calcium scores, which can be useful for reliably detecting cardiovascular risk in patients. In our pre-processing method, each lead of the ECG is segmented into several waves with an interval, which is determined as the period from the starting point of a P-wave to the end point of a T-wave. The number of segmented waves of one lead represents the number of heartbeats of the subject per 10 s. The segmented waves of one cycle are transformed into normalized waves with an amplitude of 0–1. Owing to the use of eight-lead ECG waves, the input ECG dataset has two dimensions. We used a convolutional neural network with 16 layers and 5 fully connected layers, comprising a one-dimensional filter to examine the normalized wave of one lead, rather than a two-dimensional filter to examine the coherence among the unit waves of eight leads. The training and testing are repeated 10 times with a randomly assigned dataset (177,547 ECGs). Our network model achieves an average area under the receiver operating characteristic curve of 0.801–0.890, and the average accuracy is in the range of 72.9–80.6%.

Sub-Block과 밝기 분포의 가중치를 이용한 센서스 변환 기반 스테레오 매칭 이종철 1) , 임창경 2) , 홍현기 3) Stereo matching based census transform using the sub-block and weights of the brightness distribution Jong-chul Lee 1) , Changkyoung Eem 2) , Hyunki Hong 3) 요 약