As the number of assembly parts is increasing, the time to train workers is becoming longer in cellular manufacturing. Therefore the efficient training method is required. One of the prospective methods is an instruction using motion pictures indicating how each part is combined displayed on a computer display, which may replace the current paper documents and OJT. The participants were 12 college students. The number of parts to be assembled as a time was manipulated from 1 to 6. The effect of playback speeds was also estimated on the training efficiency. Three major findings are:
(1) When slower speed of playback and less number of parts to be assemble at a time were combined, subjects' performance increased by reducing response errors.
(2) As the number of parts to be assembled at a time increased, since subjects learned the way of assembling more correctly, subjects' performance increased by reducing non response errors.
(3) As the number of parts to be assembled at a time increased, subjects learned the way of assembling faster.

View full abstract

Using a 2-dimensional body shape model, Kurokawa et al. advocated a method for describing human body shape with 15 technologically important curves, including silhouettes. In this study, essential feature points were extracted from 173 points in Kurokawa's method to create a much simpler body shape model. A total of 40 feature points, 24 points from the lateral silhouette and 16 points from the frontal silhouette, were chosen including concave and convex apexes, which were considered to contain important information for body shape analysis. Principal component analysis was performed to classify silhouette curves obtained from 428 Japanese women. It was confirmed that our method with 40 feature points was useful for describing human body shape.

View full abstract

Based on longitudinal data from 41 Japanese girls aged between 4 and 17 years (a total of 465 samples), body shape changes in the growth period were analyzed and categorized. We utilized 40 feature points in the side and front silhouettes, which were shown to be useful for describing 2-dimensional body characteristics in our previous study. Using principal component analysis of 465 samples, 5 principal components in the side silhouette and 3 principal components in the front silhouette were extracted with a cumulative contribution rate above 80%. It was also found that yearly body changes in the growth period could be categorized in 4 cluster groups.

View full abstract

[PURPOSE] We measured an accommodative response power in order to investigate the permissible limit of a binocular disparity for a visual function when gazing at stereoscopic images presented by a three dimensional display.
[METHODS] Accommodative step responses were measured using an infrared optometer, setting of a far target at the distance of 1m and a near target at the distance of 50cm. The step response of the right eye was examined 4 times at 10-second intervals. The far target was a starburst and the near target was a white circle figure on the parallax barrier system three-dimensional display as a stereoscopic image generated with, +0.5°, +1.0° and +1.5° binocular crossed disparity, and, -0.5°, -1.0° and -1.5° binocular uncrossed disparity, and 0° disparity. Spatial characteristics of the accommodative step response were quantified in terms of the accommodative response power.
[RESULTS] When the stereoscopic images generated with a+1.0° and a+1.5° binocular crossed disparities were perceived, the accommodative response waveforms were excessive accommodation (accommodation lead). In both conditions, the mean±SD values of the accommodative response powers were 1.06±0.12D and 1.26±0.16D, respectively, and exceeded the amount of accommodation stimuli.
[CONCLUSIONS] It is suggested that +1.0 degree or more crossed disparity must be avoided when displaying stereoscopic images.

View full abstract