Search results

The 3D dynamic clothing simulation is widely used in computer-added garment design. Collision detection and response are the essential component and also the efficiency bottleneck in the simulation. The purpose of this paper is to propose a high efficient collision detection algorithm for 3D clothing-human dynamic simulation to achieve both real-time and virtually real simulation effects.

The authors approach utilizes the offline data learning results to simplify the online collision detection complexity. The approach includes two stages. In the off-line stage, model triangles with most similar deformations are first, partitioned into several near-rigid-clusters. Clusters from the clothing model and the human model are matched as pairs according to the fact that they hold the potential to intersect. For each cluster, a hierarchical bounding box tree is then constructed. In the on-line stage, collision detection is checked and treated parallelly inside each cluster pairs. A multiple task allocation strategy is proposed in parallel computation to ensure efficiency.

Reasonably partitioning the 3D clothing and human model surfaces into several clusters and implementing collision detection on these cluster pairs can efficiently reduce the model primitive amounts that need be detected, consequently both improving the detection efficiency and remaining the simulation virtual effect.

The current methods only utilize the dynamic clothing-human status; the authors algorithm furthermore combines the intrinsic correspondence relationship between clothing and human clusters to efficiently shrink the detection query scope to accelerate the detection speed. Moreover, partitioning the model into several independent clusters as detection units is much more profitable for parallel computation than current methods those treat the model entirety as the unit.

Clothing retrieval is very useful to help the clients to efficiently search out the apparel they want. Currently, the mainstream clothing retrieval methods are attribute semantics based, which however are inconvenient for common clients. The purpose of this paper is to provide an easy‐to‐operate apparels retrieval mode with the authors' novel approach of clothing image similarity measurement.

The authors measure the similarity between two clothing images by computing the weighted similarities between their bundled features. Each bundled feature consists of the point features (SIFT) which are further quantified into local visual words in a maximally stable extremal region (MSER). The authors weight the importance of bundled features by the precision of SIFT quantification and local word frequency that reflects the frequency of the common visual words appeared in two bundled features. The bundled features similarity is computed from two aspects: local word frequency; and SIFTs distance matrix that records the distances between every two SIFTs in a bundled feature.

Local word frequencies improves the recognition between two bundled features with the same common visual words but different local word frequency. SIFTs distance matrix has the merits of scale invariance and rotation invariance. Experimental results show that this approach works well in the situations with large clothing deformation, background exchange and part hidden, etc. And the similarity measurement of Weight+Bundled+LWF+SDM is the best.

This paper presents an apparel retrieval mode based on local visual features, and presents a new algorithm for bundled feature matching and apparel similarity measurement.

Skeleton plays an important role in representing the essential feature of garment in image. General skeleton extraction methods often yield many short skeletal branches. Though short branches reflect the geometric details of the garment, they are obstacles in extracting the essential features. The purpose of this paper is to provide an approach to hierarchically remove them to reveal the level of details (LOD) of the skeleton, thus both the essential skeleton and the geometric skeletal branches can be definitely extracted and separated.

First, the initial garment image skeleton is extracted and smoothed. Then, the hierarchically removing mechanism is established on scoring the importance of each skeletal branch by an altered PageRank method and computing the symmetry among skeletal branches.

Experimental examples show that this method can extract and separate garment essential skeleton as well as geometric skeletal branches hierarchically. Garments in same class have a similar essential skeleton with detailed differences, so this approach can be potentially applied in garment recognition and style specification.

Traditionally, there is almost no work attempts to build LOD in skeleton of planar shapes. This paper provide an automatic device for building LOD skeleton for garment image. In another word, hierarchic skeletons with details in different prominence level are gradually established. And pairs of symmetric skeletal parts are found by taking advantage of symmetry characteristic of garment. This method is efficient in garment image skeleton extraction.

Detailed body sizes are prerequisite for made to measure or customized manufacture. Nowadays, detailed body sizes can be precisely obtained by using 3D scanners, however, the high prices of the scanners block the population for such approaches. The purpose of this paper is to provide an economical and accurate data-driven method which can predict detailed body sizes with a small number of feature sizes.

First, the representative body sizes are extracted from dozens of detail body sizes by using factor analysis and garment knowledge. Among the representative body sizes, those that are easy to be measured are selected as the feature parameters (FPs). Second, by mining the database of the body sizes, mapping from the FPs to the detailed body sizes is expressed by a combination of radial basis function and multiply linear regression. Thus, for an individual human body, his/her detailed body sizes can be predicted by a small number of FPs.

First, FPs which are easily measured and represent the main shape information of a human body are extracted. Second, detailed body sizes can be functionally predicted by the FPs.

Traditionally, measuring dozens of body sizes for each human body is tiresome and the accuracy of the sizes depends on the experience of the gaugers. In this paper, a small number of body sizes are selected as the FPs which are easy to be measured and can functionally express the other body sizes. Thus, by only measuring the FPs, the detailed body sizes can be intelligently and automatically predicted. This approach is meaningful to improve the intelligence and accuracy of the measurement, so that even an inexperienced gauger is competent to obtain accurate detailed body sizes.

Social presence (SP), which refers to individuals’ perception of others being engaged as “real people” in the same situation, is a crucial component in technology-rich learning environments (TREs). This study aims to identify major learning design, antecedents and outcomes of SP within TREs, and identify common findings from the past two decades.

Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses review principles and a qualitative analysis of selected articles, a final review of 72 studies that met inclusion criteria was obtained. Key information, including education level, discipline, sample size, study type and measurements, was extracted and studies were further analyzed and synthesized based on design features and learning modes.

The study identifies five crucial factors for instructional design to foster SP in TREs: technology affordances, multimedia features, social factors, instructional principles, learner characteristics and learning management systems. The authors compare two learning modes across three dimensions and identify popular technologies used in studies related to SP over the past two decades. Practical recommendations are provided for educators and educational technology developers to enhance SP within technology-rich learning environments.

This research contributes to the discourse on online learning and computer-supported communication, particularly in the post-COVID-19 era. By examining factors influencing SP and providing implications for instruction and educational technology development, this study offers evidence-based support to educators for engaging learners and fostering authentic learning experiences through adaptive selection of educational technologies.