Strategies for the Calculation of the Circle Hough Transform in Low-Resources Systems

In this paper, we present a method for implementing the Circle Hough Transform in systems with low computational resources, designed especially for embedded systems based on low-resource microcontrollers. Comparing with the classical approach of the transform, the proposed method allows to reduce the processing time, increase the detection precision and reduce the dimension of the Hough accumulator, which allows to reduce memory consumption. The method can be applied in different branches of Bioengineering. To quantify the results a proprietary database is used, which contains low-resolution images of human eyes, built under different lighting conditions. The results obtained show a reduction in processing time of 62%, an improvement in accuracy of 11.6% and a reduction in memory consumption.