Optores proudly introduces OGOP – the Optores GPU OCT Processing library, an ultra-fast and highly reliable processing library for optical coherence tomography (OCT) data. Based on graphic-processing unit (GPU) optimized code, OGOP enables real-time processing with sustained throughput of several gigabytes per second. The library can be licensed in editions with increasing feature sets and processing capabilities.
Unique features
- Blazing speed
- Massive parallel GPU powered OCT processing; typical throughput is over 3.5 GiB/s of sampling data on a NVidia GTX 680 GPU
- Scalability
- Multi-GPU support for multi-GVoxel/s applications
- Future-proof
- Support for OpenCL and CUDA means independence from any special hardware architecture
- Cross-platform
- Cross platform C API with strong type safety and detailed error reporting. Robust object access. C++ class bindings available
- Asynchronous exec.
- Raw data frames are queued and processed asynchronously
- Full OCT processing
-
The library implements the full Fourier-Domain processing chain, from raw data to OCT image:
- Background subtraction
- Apodization: several windowing functions available, support for custom window functions through look-up-tables
- iFFT: real to real, real to complex
- Dynamic range compression
- K-space linearization: several interpolators available for selectable tradeoff between processing throughput and quality: Nearest, Linear, Hermite, Lanczos
- Open GL integration
- Fully integrated with OpenGL for easy implementation of real-time display. Versatile support for the full range of OpenGL image formats
- Flexible in / output
- Support for all common data types: Unsigned and signed integer Single precision floating point Single precision complex floating point