To solve the problems of missing information and masking features in traditional pooling methods, a new Gaussian stochastic pooling (GS-pooling) method is proposed in this paper. The elements in the pooling domain are assigned corresponding weight values, and then the multinomial distribution sampling is sampled according to the weight values. Finally, the pooling result is obtained according to the corresponding Gaussian kernel. We conduct comparative experiments on the MNIST, CIFAR-10, and Market-1501 datasets. The results demonstrate that our pooling method achieves the best performance.
Light field display is a true three-dimensional (3D) display technology that can produce full depth cues without using any special glasses. A super multi-view display (SMV) technology is one kind of a promising method to realize the light field display, in which a large number of views with an interval smaller than the pupil diameter was produced so that two or more rays passing through one object point in a 3D space enter the pupil simultaneously. Therefore, a viewer could be able to focus on that object point. However, the main bottleneck of its commercialization is the low 3D image resolution with a small field of view, shallow depth, and et. cl. which are limited by the bandwidth of the display devices. To solve this problem, an SMV system with the slanted lenticular lens array attached to a flat panel display was developed in this article. To produce a light field display with a high resolution while maintaining two or three rays entering the pupil, an eye-tracking system is applied. The viewing zone is generated only around the viewer's eyes. The resolution is improved by the reduction of the viewing zone. The crosstalk caused by the manufacture is another issue that affects the 3D image quality. The optical model is developed to simulate the optical characteristics of light field display based on the raytracing considering the fabrication error. The calibration method is proposed to compensate for the error and improve the image quality further.
A super multi-view display (SMV) technology is one kind of a promising method to realize the light field display approximately. It generates a dense viewing points with an interval smaller than the pupil diameter so that two or more rays passing through one object point in a three-dimensional (3D) space enter the pupil simultaneously. Therefore, a viewer could be able to focus on that object point. To produce a light field display with high 3D resolution while maintaining two or three rays entering the pupil, a flat panel system with high resolution or multiple display devices are required. To solve this problem, an SMV system with the slanted lenticular lens attached to a liquid crystal display (LCD) was developed in this article. The viewing zone is generated only around the viewer's eyes. Therefore, the resolution is improved by the reduction of the viewing zone. However, the viewing freedom is limited not only in the horizontal and vertical direction but also in the longitudinal direction. To increase the viewing freedom, an eye-tracking system is applied to the developed SMV system. The dynamical viewing zone characterizations including viewing interval, spatial distribution, and et. cl in this system was analyzed. To keep the viewer’s eyes always in the viewing zone with the smaller crosstalk, the compensation factors were calculated to generate the viewing zone depending on the 3D positions of the viewer’s eyes. The results show that the proposed SMV system successfully increased the horizontal, vertical, and longitudinal viewing freedom.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.