Skeleton Figure Lab Report

III. MEHOD When we are designing pose using existing software, prepared 3DCG character's joint onto computer, which we will design a pose while operating using mouse little by little. This is also many very time-consuming work effort. As this research, we use skeleton figure for posture design. The skeleton figure because the user can directly touch own hands, it is possible to intuitively poses design. After pose design, to take a picture of it. This photo become the input image. And this pose is applied to the 3DCG model on the computer. From input to results images have shown in the Fig.1. Next show the estimation procedure of the human body posture below.

1) Input photos of the skeleton figure wearing the markers

2) Identification of the markers in the image

3) Assignment of the positional relationship of coordinates and joint of each marker

4) Calculation of the slope of the attitude

5) Output of result as the pose data

We attach the markers to joints of the skeleton figure of 14 joints (waist, neck, shoulders, elbows, hands, groins, knees, feet). In order to identify the respective body parts were prepared marker 3 colors. Waist and neck green, arms red, legs blue. Background color eliminate overlap with the color markers and the object. To identify the skeleton figure and markers using HSV color diagram. From markers that are color-corded

Then, we measure the concentration of shading between adjacent joint and joint. In other words, the slope of each portion, that is to perform the depth estimation. We assume that is above the subject at the light source. Position and its relationship to each joint was determined by the method “A, B and C”. In the same way as the estimate of the elbow or knee joint, we connect a line between the adjacent joint and joint. To measure the concentration of shadow of this straight line. That is estimated the slope of each site to the depth