3D Perceptions Research Articles

3D printing exposure and perception in radiology residency: survey results of radiology chief residents

Rationale and objectivesThe purpose of this study is to summarize a survey of radiology chief residents focused on 3D printing in radiology.Materials and methodsAn online survey was distributed to chief residents in North American radiology residencies by subgroups of the Association of University Radiologists. The survey included a subset of questions focused on the clinical use of 3D printing and perceptions of the role of 3D printing and radiology. Respondents were asked to define the role of 3D printing at their institution and asked about the potential role of clinical 3D printing in radiology and radiology residencies.Results152 individual responses from 90 programs were provided, with a 46% overall program response rate (n = 90/194 radiology residencies). Most programs had 3D printing at their institution (60%; n = 54/90 programs). Among the institutions that perform 3D printing, 33% (n = 18/54) have structured opportunities for resident contribution. Most residents (60%; n = 91/152 respondents) feel they would benefit from 3D printing exposure or educational material. 56% of residents (n = 84/151) believed clinical 3D printing should be centered in radiology departments. 22% of residents (n = 34/151) believed it would increase communication and improve relationships between radiology and surgery colleagues. A minority (5%; 7/151) believe 3D printing is too costly, time-consuming, or outside a radiologist’s scope of practice.ConclusionsA majority of surveyed chief residents in accredited radiology residencies believe they would benefit from exposure to 3D printing in residency. 3D printing education and integration would be a valuable addition to current radiology residency program curricula.

The Perception of the Vertical Dimension (3D) through the Lens of Different Stakeholders in the Property Market of China

China has experienced fast urbanization with a growing urban population, which has inevitably led to the adoption of a vertical housing style with high-rise buildings. However, how people subjectively perceive the vertical dimension (3D) in the property market is neither adequately documented nor well understood. The 3D perception helps us to understand a myriad of social and psychological effects of living in high-rise buildings. We organized and conducted semi-structured expert interviews, focus groups, and the circulation and compilation of questionnaires in Xi’an, China, to investigate how different stakeholders in the property market perceive 3D. The results show that: (1) real estate developers do not adjust property prices for specific 3D factors, and the local government does not consider 3D in housing policies; (2) the current status of 3D modeling in Xi’an is still in the embryonic state; (3) 3D factors are highly valued by buyers but not well-understood by real estate developers and local government. In addition, 3D factors score higher than horizontal (2D) factors (1.12 to 0.88). Gender and age groups do not influence housing preferences concerning 2D and 3D factors. These findings provide valuable insights for real estate developers concerning pricing policies and the local government concerning housing policies. In the future, 3D perceptions and factors should be prioritized in order to improve urban infrastructure and ensure the increased availability of, and fair public access to welfare related to 3D in urban areas.

Volumetric Propagation Network: Stereo-LiDAR Fusion for Long-Range Depth Estimation

Stereo-LiDAR fusion is a promising task in that we can utilize two different types of 3D perceptions for practical usage - dense 3D information (stereo cameras) and highly-accurate sparse point clouds (LiDAR). However, due to their different modalities and structures, the method of aligning sensor data is the key for successful sensor fusion. To this end, we propose a geometry-aware stereo-LiDAR fusion network for long-range depth estimation, called volumetric propagation network. The key idea of our network is to exploit sparse and accurate point clouds as a cue for guiding correspondences of stereo images in a unified 3D volume space. Unlike existing fusion strategies, we directly embed point clouds into the volume, which enables us to propagate valid information into nearby voxels in the volume, and to reduce the uncertainty of correspondences. Thus, it allows us to fuse two different input modalities seamlessly and regress a long-range depth map. Our fusion is further enhanced by a newly proposed feature extraction layer for point clouds guided by images: FusionConv. FusionConv extracts point cloud features that consider both semantic (2D image domain) and geometric (3D domain) relations and aid fusion at the volume. Our network achieves state-of-the-art performance on KITTI and Virtual-KITTI datasets among recent stereo-LiDAR fusion methods.

Perception, Cognition, and Action in Hyperspaces: Implications on Brain Plasticity, Learning, and Cognition.

We live in a three-dimensional (3D) spatial world; however, our retinas receive a pair of 2D projections of the 3D environment. By using multiple cues, such as disparity, motion parallax, perspective, our brains can construct 3D representations of the world from the 2D projections on our retinas. These 3D representations underlie our 3D perceptions of the world and are mapped into our motor systems to generate accurate sensorimotor behaviors. Three-dimensional perceptual and sensorimotor capabilities emerge during development: the physiology of the growing baby changes hence necessitating an ongoing re-adaptation of the mapping between 3D sensory representations and the motor coordinates. This adaptation continues in adulthood and is quite general to successfully deal with joint-space changes (longer arms due to growth), skull and eye size changes (and still being able of accurate eye movements), etc. A fundamental question is whether our brains are inherently limited to 3D representations of the environment because we are living in a 3D world, or alternatively, our brains may have the inherent capability and plasticity of representing arbitrary dimensions; however, 3D representations emerge from the fact that our development and learning take place in a 3D world. Here, we review research related to inherent capabilities and limitations of brain plasticity in terms of its spatial representations and discuss whether with appropriate training, humans can build perceptual and sensorimotor representations of spatial 4D environments, and how the presence or lack of ability of a solid and direct 4D representation can reveal underlying neural representations of space.

3D object detection with deep learning

Finding an appropriate environment representation is a crucial problem in robotics. 3D data has been recently used thanks to the advent of low cost RGB-D cameras. We propose a new way to represent a 3D map based on the information provided by an expert. Namely, the expert is the output of a Convolutional Neural Network trained with deep learning techniques. Relying on such information, we propose the generation of 3D maps using individual semantic labels, which are associated with environment objects or semantic labels. So, for each label we are provided with a partial 3D map whose data belong to the 3D perceptions, namely point clouds, which have an associated probability above a given threshold. The final map is obtained by registering and merging all these partial maps. The use of semantic labels provide us a with way to build the map while recognizing objects.

Factors that determine depth perception of trapezoids, windsurfers, runways.

We report here a windsurfer1 illusion, a naturally occurring trapezoidal illusion in which the small end of the sail viewed at a distance appears to be pointed away from the observer even when it is closer. This naturally occurring illusion is so compelling that observers are unaware of their gross perceptual misinterpretation of the scene. Four laboratory experiment of this kind of trapezoidal illusion investigated the joint effects of retinal orientation, head position, relative motion, and the relative direction of gravity on automatic depth perception. Observers viewed two adjacent white trapezoids outlined on a black background rotating back and forth ± 20° on a vertical axis much like the sails of two adjacent windsurfers. Observers reported which side of the trapezoids (long or short) appeared to be closer to them (i.e., in front). The longer edge of the trapezoid was reported in front 76 ± 2% of trials (“windsurfer effect”) whether it was on the left or on the right. When the display was rotated 90°to produce a runway configuration, there was a striking asymmetry: the long edge was perceived to be in front 97% when it was on the bottom but only 43% when it was on top (“runway effect”). The runway effect persisted when the head was tilted 90° or when displays on the ceiling were viewed from the floor. Ninety-five percent of the variance of the variance in the strikingly different 3D perceptions produced by the same 2D trapezoid image was quantitatively explained by a model that assumes there are just three additive bias factors that account for perceiving an edge as closer: Implicit linear perspective, lower position on the retina (based on an automatic assumption of viewing from above), and being lower in world coordinates.

A Frontal Multi-Projection Autostereoscopic 3D Display Based on a 3D-Image-Guided Screen

A frontal multi-projection (FMP) autostereoscopic display using a 3D-image-guided screen is proposed for improving space efficiency. Based on the optical multiplexing and de-multiplexing, the important optical functions of 3D-image-guided screen are parallax image interlacing and view-separating, which is capable of reconstructing 3D images without quality degradation from the front direction. The operating principle, optical design calculation equations and correction method of parallax images are described in detail. A 50-inch FMP autostereoscopic prototype based on the 3D-guided-image screen with pixels $({\hbox{1024}}\times {\hbox{768}}\times {\hbox{8}})$ is developed, which enhances the brightness and 3D perceptions. The performance of this prototype is evaluated by measuring the luminance and crosstalk distribution along the horizontal direction at the optimum viewing distance. This research has potential applications in some commercial entertainments and movies for the realistic 3D perceptions.

비대칭적 3차원 영상에 대한 인간의 인지 특성

다시점 비디오의 일반적인 형태인 3차원 방송 서비스는 최근 많은 관심을 받고 있다. 그렇지만 3차원 방송의 성공을 위해서는 상용 가능한 컨텐츠의 개발, 사용의 용이성, 방송 품질 개선, 비용 절감 등의 과제가 남아있다. 본 논문은 이러한 과제들 중 방송 품질 개선 측면, 특히 인간 요소를 고려한 주관적 화질 개선을 중심으로 접근하여 양안식 영상과 다시점 영상에 적합한 비대칭 코딩 방법을 설명하고 최적의 주관적 화질을 유지하는 비대칭 율의 정량적인 값을 제안한다. 또한 저화질의 2차원 영상들의 경계선들이 3차원 영상의 인지에 미치는 영향을 실험적으로 분석하고 경계선 왜곡의 문제점을 주관적 화질 측면에서 개선하는 경계선 보호 알고리듬을 제안한다. 실험 결과를 통해 기존의 부호화 방법보다 주관적 화질을 개선시킴을 확인하였다. The 3DTV services can be seen as a general case of the multi-view video that has been receiving a significant attention lately. However, the key factors that influence the success of 3DTV are the availability of content, the ease of use, the quality of contents, and the reduction of cost. This paper deals primarily with the perceptual improvement in image quality, especially based on human factors. An optimal asymmetrical coding method for binocular and multi-view images is presented. The quantitative value of asymmetrical rate to maintain optimized subjective image quality is explored. Also we analyze how edges of 2D images affect on 3D perceptions and propose an edge-preserving algorithm to perform perceptual improvements. Experimental results demonstrate that the proposed algorithm enhances subjective image quality much better than conventional methods.