The new year has come and we are already at work to improve further the core technology behind 3DF Zephyr.
I opened this thread to give you a brief insight of what you should expect with the next updates of Zephyr.
We are now internally testing the version 1.005, which will be released during the second part of the month. The new release contains an important technology upgrade of the structure from motion pipeline (Samantha). The bundle adjustment routines and some geometric data structures have been rewritten; this lead to improved accuracy, improved running times, and a reduced memory footprint. The improvements are more remarkable when dealing with very large datasets.
In February we will be releasing an upgrade for our multiview stereo algorithm. Everything will be executed in CUDA, so expect a big speedup. We are also adding the support for multiple GPUs for both the structure from motion and stereo parts and improvements on the texture generation. Both this improvements will be contained in the February upgrade.
More to come, happy new year.
I opened this thread to give you a brief insight of what you should expect with the next updates of Zephyr.
We are now internally testing the version 1.005, which will be released during the second part of the month. The new release contains an important technology upgrade of the structure from motion pipeline (Samantha). The bundle adjustment routines and some geometric data structures have been rewritten; this lead to improved accuracy, improved running times, and a reduced memory footprint. The improvements are more remarkable when dealing with very large datasets.
In February we will be releasing an upgrade for our multiview stereo algorithm. Everything will be executed in CUDA, so expect a big speedup. We are also adding the support for multiple GPUs for both the structure from motion and stereo parts and improvements on the texture generation. Both this improvements will be contained in the February upgrade.
More to come, happy new year.
Comment