Assessment of hardware vs software implementations for video microscopy

Video microscopic platforms have been developed to capture cell dynamics, such as cell velocity, in real-time to quantify and usefully leverage the electro kinetic response of cells to electric fields. The first and critical task in a cell motion detection algorithm is the computationally demanding segmentation process, where a static cell boundary aids grouping the cell regions and thereby identifying the cell. We propose the use of dedicated hardware to increase the speed and efficiency of operation and the computational effort involved in the segmentation process. This dedicated hardware is compared to a software program executed on a general purpose processor to ascertain the relevance of the proposed hardware. The hardware design is verified on a FPGA to implement the cell segmentation architecture, modelled in Verilog and synthesized using QuartusII Altera Stratix device. The segmentation results presented shows relevant design statistics demonstrating equivalent cell identification performance, however, with a computational speed-up of ~5 times compared to the software based approach.