Research On 3D Garment Modeling Based On Human Feature Recognition

Garment is not only an important symbol of the social culture, but also the necessity of one's daily life. Recently, with the rapid development of economy and society, garment is increasing characterized by the personalization, fitness and fashion as well. However,traditional custom tailoring is an approach of single-volume and single-cut, which is subject to the tailor's working experience. Due to the limitation of long production cycle, poor product quality stability, and large fitting error, traditional custom tailoring is currently difficult to meet the needs of existing mass customerization.Focusing on some key technique problems of mass customerization in the garment industry, human body characteristics recognition and 3D garment prototyping are investigated in this dissertation. Based on the propotional relationship of characteristic points regarding the human body, random forest classification algorithm is applied to the statistical inference for the character of human body. In view of recognized characteristic points, a subdivision prototyping algorithm of the 3D garment is proposed by dividing the overall 3D garment pattern into a number of developable patches such as the shoulder patch, the chest patch, the abdominal patch and the transitional patch. For taking the developability requirements into account, flattening of the 3D garment patch is implemented through elementary geometry operations. Subsequently the hierarchical B-spline is applied to shape control and detail editing of the designated 3D patch to fit the human shape preferably. To model the mechanical properties,i.e. bending, shear and tension inherent to fabric, a spring-mass model is introduced to simulate the dressing behavior of 3D garment prototype by incorporating Euler's numerical solution method and axis aligned bounding box based collision detection and response mechanism.Experiment results show that the proposed methods in this dissertation are efficient and feasible in the promotion to mass customerization garment products. It is consequently anticipated to reveal prospective applications in the garment industry.