The relationship between the built environment and urban street vitality, as a key issue of contemporary design, has been discussed over decades. However, most existing studies relying on linear regression models do not reveal complicated impacts features often neglect their threshold effects. As response, this study applies gradient boosting decision tree (GBDT) model with large amount new data to explore in-depth understanding vitality. Based samples from 12 Chinese cities, series...

In this paper, we propose a technique for automatically annotating visualizations according to the textual description. our approach, visual elements in target visualization, along with their properties, are identified and extracted Mask R-CNN model. Meanwhile, description is parsed generate search requests. Based on identification results requests, each descriptive sentence displayed beside described focal areas as annotations. Different sentences presented various scenes of generated...

In this paper, we propose a novel approach to generate captions for visualization charts automatically. the proposed method, visual marks and channels, together with associated text information in original charts, are first extracted identified multilayer perceptron classifier. Meanwhile, data can also be retrieved by parsing mapping relationships. Then 1-D convolutional residual network is employed analyze relationship between elements, recognize significant features of both as input. final...

We propose an automatic pipeline to generate visualization with annotations answer natural-language questions raised by the public on tabular data. With a pre-trained language representation model, input natural and table headers are first encoded into vectors. According these vectors, multi-task end-to-end deep neural network extracts related data areas corresponding aggregation type. present result carefully designed for different attribute types tasks. conducted comparison experiment...

The human skeleton joints captured by RGB-D camera are widely used in action recognition for its robust and comprehensive 3D information. Presently, most methods based on treat all skeletal with the same importance spatially temporally. However, contributions of vary significantly. Hence, a GL-LSTM+Diff model is proposed to improve actions. A global spatial attention (GSA) express different weights provide precise information recognition. accumulative learning curve (ALC) introduced...

Precise identification of blister packages carries utmost importance at dispensing stations, where numerous prescriptions are to be efficiently dispensed by pharmacists. However, the usual presence several hundreds similarly looking, but completely different types packages, in a crowded station makes it prone human error, posing serious safety and health concerns for patients life. In this work, we propose highlighted deep learning (HDL) based approach accurate packages. HDL allows smart...

Human action recognition from video sequences is one of the most challenging computer vision applications, primarily owing to intrinsic variations in lighting, pose, occlusions, and other factors. The human skeleton joints extracted by depth camera Kinect have advantages simplified structures rich contents, are therefore widely used for capturing actions. However, at present, skeletal joint Deep learning based methods treat all equally both spatial temporal dimensions. Logically, this not...

Prescription dispensing accuracy is of paramount importance for all hospitals. However, human errors are inevitable due to multiple reasons, such as fatigue, stress, heavy workload, lack effective verification measures, mismanagement. Such pose serious safety and health concerns on the part patients may well lead a series medical disputes. Based induced deep learning, this paper proposes real-time Blister Package Identification System (BPIS) assist pharmacists' drug dispensing. Under...

Success in delivering dynamic urban coastal zones is considered essential, as it brings enormous opportunities to the social, economic, ecological, and cultural development of cities addition benefitting zones. However, environmental drivers contributing zone vitality remain uncertain due unclarified spatial boundaries influences diverse characteristics from surrounding contexts. This study aims at exploring that can vitalize inform an effective way instruct design procedures. It sets out...

Combining depth information and color image, RGB-D cameras provide ready detection of humans the associated 3D skeleton joints data, facilitating if not revolutionizing conventional image centric researches in, among others, computer vision, surveillance, human activity analysis. Applicability a D-RBG camera, however, is restricted by its limited range frustum in 0.8 to 4 meters. Although camera network, constructed deployment several at various locations, could extend coverage, it requires...

Good action recognition relies on correct interpretation of two critical attributes related to action: the spatial attribute detected person's posture, and temporal body movement. Whereas deep learning has greatly improved image recognition, we have not found a similar progress for recognition. One main reasons is due complexity caused by additional dimension; another, fact that there are less annotated training data samples than In this regard, paper proposes handcrafted cued LSTM model...

Creating thematic sessions based on accepted papers is important to the success of a conference. Facing large number from multiple topics, conference organizers need identify topics and group them into by considering constraints session numbers paper in individual sessions. In this paper, we present system using visualization topic modeling help construction The provides automatically generated schemes allows users create, evaluate, manipulate with given constraints. A case study our VAST...

Viewpoint variation has been a major challenge in dealing with comparison-based image processing. Reduction or total removal of viewpoint the common pursuit many processing applications such as human motion analysis and gesture analysis. By exploiting three-dimensional (3-D) skeletal joints information provided by RGB-D cameras Kinect, this study proposes skeleton-based invariant transformation (SVIT) technique that transforms 3-D skeleton data into an orthogonal coordinate system...

One main difficulty in applying action recognition to practical applications is the need segment beginnings and ends of actions a continuous online monitoring process. This paper proposed finite state machine (FSM) model for automatic segmentation solution, based on pose streams form skeleton joint data provided by Kinect. With problem reframed as identification problem, key solution hinges detection changing events, which signify start new underlying action. In that regard, decision tree...

Combining depth information and color image, D-RGB cameras provide a ready detection of human associated 3D skeleton joints data, facilitating, if not revolutionizing, conventional image centric researches in, among others, computer vision, surveillance, activity analysis. Applicability D-RBG camera, however, is restricted by its limited range frustum in the 0.8 to 4 meters. Although camera network, constructed deployment several at various locations, could extend coverage, it requires...

Kinect have been used as a revolutionary sensor for recent human activity recognition research, mainly due to its ready skeletal joint information that facilitates analysis. However, the sensor's unstable and imprecise measurement may impair analysis results. To alleviate this impact unavoidable noisy measurement, paper presents approach based on relative positional relationship among measured joints data. Relative between two refers relation either one joint's position in x, y, z...

In this paper, we propose EathquakeAware, a visual analytics system for social media visualization. Our supports analyzing geo-tagged messages from different perspectives, including location, keyword, account. It can summarize the evolution patterns of topics and reveal discussing trend keywords to help evaluate situations in regions. We take VAST Challenge 2019 Mini 3 as case studies demonstrate its effectiveness fictitious analysis scenario.

Karada OK, a term coined to bear similarity the popular Karaoke, is body gesture matching game. In time span of song, video movement performed by an exemplary coach recorded as template, later, players are match gestures along with music, score given at end Due unavoidable camera view difference between one recording template and that game playing, direct comparison for scoring can be difficult. This paper presents Kinect-based solution which utilizes Skeleton-based Invariant Transformation...