Real-time holographic rendering of medical imaging
On the use of Coaxlink technology to overcome the limitations of real-time observation of fundus vasculature by holographic laser Doppler imaging in clinical ophthalmology.
The advent of very high-speed digital cameras using the CoaXPress image transmission standard now allows the collection of digital images compatible with high-frequency laser Doppler measurement. It achieves an unparalleled level of performance and image quality for non-invasive imaging of blood flow in retinal vessels.
In this example implementation, a high-speed, high-resolution, state-of-the-art camera captures both the details of the fundus vasculature and its dynamics, revealing the heartbeat.
The morphological as well as temporal details are important in order to reconstruct in real time a precise holographic image of the in-situ vasculature and its possible pathologies.
The ever-increasing size and speed of optical sensors can now capture a 512×512 pixel (12-bit) image at a frame rate over 20,000 frames per second with a Phantom S710 camera from Vision Research, a division of Ametek. Additionally, signal processing capabilities are greatly increased with the advent of graphics processing units (GPUs). In this situation, the real-time transmission of digital data at very high speed between the camera and the processor becomes a critical element.
To achieve this digital data transfer capability, two parameters can be exploited. The maximum connection speed and the number of connections implemented in parallel while guaranteeing the integrity of the data as well as the conservation of their timing. To do this, the CoaXPress industrial vision standard was naturally chosen for its ability to manage the transfer of images at very high speed in real time without any loss of information.
Open image in new tab
Open image in new tab
Open image in new tab
Holographic Doppler image constructed by digital computationHolographic Doppler image constructed by digital computation
Holovibes
It is with this in mind that the Langevin Institute, in collaboration with the Quinze-Vingt hospital in Paris have selected the Coaxlink Octo card from Euresys, in combination with the Phantom S710 camera. Offering eight CoaXPress CXP-6 connectors each, it enabled the design of the new clinical holographic Doppler imaging device, whose image acquisition and rendering software, Holovibes, is developed by the Digital Holography Foundation.
Providing comprehensive system design services, Stemmer Imaging, distributor of Euresys, assisted with the selection of the best machine vision components meeting the specific user requirements. Technical support and consulting was then provided throughout the project, ensuring the system seamless implementation and optimal performance.
The Euresys Coaxlink Octo acquisition card, compatible with the CoaXPress standard like all models in the Euresys Coaxlink series, can operate in synchronization when two or more cards are installed in the same computer. They guarantee the maximum transfer capacity of the Phantom S710 camera, for any image size.
It is used here in pairs within the same computer to reach the frame rate made available by the Phantom S710 camera through its 16 CoaXPress CXP-6 communication ports. The two Euresys Coaxlink Octo cards manage synchronized acquisition data and reconstruct in real time a single global image which is made available in the memory of the computer without delay, by direct access to the memory (Direct Memory Access – DMA). This process allows the Holovibes application to process the acquired laser interference patterns in real time to build and display holographic laser Doppler images of exceptional quality, with minimal latency.
This unprecedented performance, made possible today by the Coaxlink technology, marks an important step in ophthalmic imaging for the ultra-fast holographic Doppler laser. Like other medical imaging techniques, such as Magnetic Resonance Imaging (MRI), Computed Tomography (CT) and Ultrasound Imaging, it is expected that these improvements based on high-speed digital computing for image rendering will represent significant diagnostic advancements.