Y.L. Kwok

A robust adaptive clustering analysis method for automatic identification of clusters

Identifying the optimal cluster number and generating reliable clustering results are necessary but challenging tasks in cluster analysis. The effectiveness of clustering analysis relies not only on the assumption of cluster number but also on the clustering algorithm employed. This paper proposes a new clustering analysis method that identifies the desired cluster number and produces, at the same time, reliable clustering solutions. It first obtains many clustering results from a specific algorithm, such as Fuzzy C-Means (FCM), and then integrates these different results as a judgement matrix. An iterative graph-partitioning process is implemented to identify the desired cluster number and the final result. The proposed method is a robust approach as it is demonstrated its effectiveness in clustering 2D data sets and multi-dimensional real-world data sets of different shapes. The method is compared with cluster validity analysis and other methods such as spectral clustering and cluster ensemble methods. The method is also shown efficient in mesh segmentation applications. The proposed method is also adaptive because it not only works with the FCM algorithm but also other clustering methods like the k-means algorithm.

Block pattern generation: From parameterizing human bodies to fit feature-aligned and flattenable 3D garments

Research on clothing related CAD is blooming rapidly in the last two decades. It speeds up the product development process significantly and shortens the time to market of fashion products. Although many important results have been obtained, particularly in the computer graphics community, the textile industry is somehow reluctance to adopt these results in actual apparel manufacturing. The main concern is the accuracy of the resulted patterns, because the pattern generation processes ignored some important textile material and manufacturing constraints. This paper introduces a method for generating 2D block patterns from 3D scanned body. A parameterization process is first conducted on a scanned body to create a parameterized model, represented by horizontal B-spline curves. A basic wire-frame aligned with body features is then established based on the parameterized model. Proper clothing ease is carefully incorporated into the model by scaling the wireframe to accomplish the desired fit. Based on the deformed wireframe, a 3D flattenable garment is modeled by boundary triangulation. The main contribution of the proposed method is that the created 3D garment blocks are geometrically flattenable to produce accurate 2D patterns with optimized ease distribution to ensure garment fit. The proposed method is validated and compared to two conventional block patternmaking methods. The experimental results indicate that the proposed method is easy to implement and can generate patterns with satisfactory fit. Furthermore, the method can be used to create fit-ensured mass-customized apparel product.

Body measurements of Chinese males in dynamic postures and application

It is generally accepted that there is a relationship between body dimensions, body movement and clothing wearing ease design, and yet previous research in this area has been neither sufficient nor systematic. This paper proposes a method to measure the human body in the static state and in 17 dynamic postures, so as to understand dimensional changes of different body parts during dynamic movements. Experimental work is carried out to collect 30 measurements of 10 male Chinese subjects in both static and dynamic states. Factor analysis is used to analyse body measurement data in a static state, and such key measurements describe the characteristics of different body figures. Moreover, one-way ANOVA is used to analyse how dynamic postures affect these key body measurements. Finally, an application of the research results is suggested: a dynamic block patternmaking method for high-performance clothing design.

An efficient human model customization method based on orthogonal-view monocular photos

Human body modeling is a central task in computer graphics. In this paper, we propose an intelligent model customization method, in which customers detailed geometric characteristics can be reconstructed using limited size features extracted from the customers orthogonal-view photos. To realize model customization, we first propose a comprehensive shape representation to describe the geometrical shape characteristics of a human body. The shape representation has a layered structure and corresponds to important feature curves that define clothing size. Next, we identify and model a novel relationship model between 2D size features and 3D shape features for each cross-section using real subject scanned data. We predict a customers cross-sectional 3D shape based on size features extracted from the customers photos, and then we reconstruct the customers shape representation using predicted cross-sections. We develop a new deformation algorithm that deforms a template model into a customized shape using the reconstructed 3D shape representation. A total of 30 subjects, male and female, with varied body shapes have been recruited to verify the model customization method. The customized models show high degree of resemblance of the subjects, with accurate body sizes; the accuracy of the models is comparable to scan. It shows that the method is a feasible and efficient solution for human model customization that fulfills the specific needs of the clothing industry.

Automatic Block Pattern Generation from a 3D Unstructured Point Cloud

Accurate and fitted garment patterns are fundamentally important in garment manufacturing. Even though a virtual body can now be obtained by 3D scanning, the problem of generating patterns model is still challenging because the mapping from a 3D body to 2D pattern is constrained by complex garment style information and sewing definitions. This paper presents a new approach for generating 2D block patterns directly from scanned 3D unstructured points of the human body. The new approach consists of a series of steps from body recognition, body modelling to pattern formation. In the paper, algorithms for body feature extraction and body modelling are first described, then the relationship between the human body, patterns and darts are investigated, and pattern creation through automatic dart transformation are thus developed. The paper has demonstrated that the proposed method can generate 2D block patterns from a 3D unstructured point cloud.

A Parameter Free Approach for Clustering Analysis

In the paper, we propose a novel parameter free approach for clustering analysis. The approach needs not to make assumptions or define parameters on the cluster number or the results, while the clustered results are visually verified and approved by experimental work. For simplicity, this paper demonstrates the idea using Fuzzy C-Means (FCMs) clustering method, but the proposed open framework allows easy integration with other clustering methods. The method-independent framework generates optimal clustering results and avoids intrinsic biases from individual clustering methods.