Christian Perwaß, Raytrix

Talk Synopsis
A camera design that has regained much popularity in the last couple of years is that of the plenoptic camera where a microlens array is placed in front of the image sensor. We present our plenoptic technology with a high effective resolution of up to one quarter of the sensor resolution, an extended depth of field and fast CUDA implementations of lightfield algorithms that make live 3D-video possible. The microlens technology developed by Raytrix increases the depth of field through specially designed microlens arrays that consist of microlenses of different focal lengths. A Raytrix camera therefore acts like a number of interlaced plenoptic cameras focused to different depths, so that refocusing and depth estimation are available over the whole extended depth of field.

The Raytrix camera system currently available is capable of recording, processing and displaying 3D video with up to 3 megapixels effective resolution at up to 6fps using CUDA implementations of its own algorithms running on an NVIDIA GTX580 graphics card. As the recorded video consists of lightfield images, the possibilities of post-production are extended to pixel-wise focusing over an extended depth of field, variable stereo base-line, stereo base-line rotation and slight view point changes.

We use cookies in this website, by continuing using this website you agree to us using them.


The British Machine Vision Association and Society for Pattern Recognition