The measure and estimation of human foot volume using 3D range data

• Main Authors: Lu,BO-Sheng, 呂柏昇
• Other Authors: Ching-Hua Lin
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/sqn3f5

碩士 === 中華大學 === 工業管理學系碩士班 === 101 === Human foot volume is one of the significant physical characteristics in anthropometry. It is widely applied as a parameter and prognosis index in biomechanics and clinical therapies respectively. The measurement of foot volume has been established including water volumetry, measuring from dissected corpses, reaction force plate, and etc. CT (Computed Tomography) scan or MRI (magnetic resonance imaging) has improved the measuring precision and accuracy relatively while it is invasive, time-consuming and requiring sophisticated equipment and associated cost. Now, the 3D scanning technology has been developed and has been becoming a master measurement in anthropometry due to its high-level of precision and accuracy, noninvasive, brief measuring time, and etc. This study was aimed to measure human foot volume using 3D scanning range data, to investigate the relationships among foot length, width, height and ankle circumference, to normalize foot volume, to derive foot volume estimator both by regression and BPNN (Back-Propagation Neural Network) and to validate and compare the two estimators. The mean foot volumes were 930.10 cm3, 626.79 cm3 and 766.94 cm3 for male, female and total samples respectively. The results showed the statistically significant difference of mean foot volume between dominant and non-dominant foot tested by both sixty-seven male samples and seventy-eight female samples. There were 101 right-dominant subjects, 70% of the totle 145 subjects. Averagely, the foot length, width and height were about thirteen, six and four percent of stature height respectively. These percentages were significantly different between two genders, right- and left-dominant groups, and among different BMI (Body Mass Index) ranks. The ratios of foot length to width, length to height and width to height were also significantly different between two genders, right- and left-dominant groups, and among different BMI ranks. This indicated that the foot shapes were slightly different between these groups. The foot volume normalized by body weight was significant different between different waist-girh ranks and under the interaction of genders*dominances*BMIs. Additionally, the normalization by foot length was significant different between different waist-girh ranks and under the interactions of genders*dominances and dominances *BMIs. Finally, the estimation of foot volume by BPNN estimator was better than by regression estimator, the estimation result by which was still acceptably accurate. All these results were elaborated and hopefully to contribute to the applications of biomechanics, clinical therapies and other disciplines.