Poster presentation | Open | Published:
A method for improvement of data quality from OPS/SDF devices in visualization of the microcirculation
Critical Carevolume 10, Article number: P317 (2006)
OPS imaging (Cytometrics, Philadelphia, PA, USA) and SDF imaging (MicroVision Medical, Amsterdam, The Netherlands) are valuable tools for visualization and study of the microcirculation. Unfortunately, the widespread current practice of using a (digital) video recorder and subsequent transfer to a computer results in data loss. This is due to video compression and nonoptimal use of the available dynamic range. In order to be able to develop and use software for automated data analysis we should strive for the highest possible image quality. Here we describe a simple and cost-effective method for obtaining significant improvement of data quality.
A 10-bit frame-grabber board (NI PCI-1409; National Instruments, Austin, TX, USA) was used for online capture of noncompressed video data from an SDF device. The 10-bit data was real-time converted to optimized 8-bit data by a dynamic scaling algorithm, compressing the excess 1024 gray values into the standard 256 gray values and thereby mathematically correcting for the offset. The converted data were written to hard disk as an uncompressed AVI file. The data stream of 2.5 Gb/min was handled by a 3 MHz Pentium-4 PC with two 250-Gbyte hard disks in the Raid 1 mode. Software was written in LabView (version 7.1; National Instruments).
The system routinely delivers excellent dynamic range and resolving power in a clinical setting. Figure 1 is an example from a sublingual measurement in a healthy volunteer. The white bar equals a distance of 100 mm, the right panel is a magnification the marked area and the arrow indicates single erythrocytes within a capillary.
A major improvement in the dynamic range (typically a factor 2) can be obtained by 10-bit data acquisition followed by 8-bit conversion by a dynamic scaling algorithm. Furthermore, noncompressed video data offer a better image quality in terms of resolving power of single erythrocytes and extremely small vessels. It is expected that the improved data quality will ease automatic analysis by future software.