This filtering is performed by first iterating over all the lines and obtaining the ρ and θ values from application
of Hough transformation and determine whether the line is horizontal or vertical. All lines with values of sin θ and cos θ
less than 0.001 are ignored and mark the rest of the lines as horizontal if sinθ < 0.001 and vertical if cosθ < 0.001. Based on the grid layout, the border lines of the erythrocyte grid are obtained. For obtaining the automatic count, at first contrast on the red channel for background elimination, then, analyze the image pixel statistics and based on this statistics find the threshold, followed by tracking of contours of the image. Next, unwanted items are eliminated using values of image statistics. This is followed by image segmentation and counting. But, in this method, no special attention is given to overlapping cells or cells with holes [7]. There exists another method for automated blood cell counting that utilizes digital
image processing for obtaining the RBC count. In this process, the first step is to delete the incomplete blood cells that overlap
on the boundary of the image. After this, single blood cell
is extracted from the image using edge detection algorithm.
Finally, the single blood cell images are analyzed by using
neural networks to count red blood cells