Improved method and apparatus for imaging of cells and finding cells, counting cells and determining confluence of a cell culture

Abstract

A method and apparatus for counting cells in image regions of a phase image of a cell culture in a cell culture vessel, has an imager for imaging a cell culture in a cell culture vessel and at least on processor for providing a phase image of the cell culture and configured to: find the local maximum of each source image region, wherein the size of each source image region is defined by an approximated diameter by creating an intermediate image using an elliptical structuring element mask having a deleted central pixel and performing a morphological operation of dilation on the source image; for each source image region create a new mask image as a logical ‘LARGER THAN’ of the intermediate image with the source image resulting in zero everywhere except at a single pixel dot located at a pixel location wherein the source image has a local maximum value of a local region and wherein the size of the local region is determined by the size of the elliptical structuring element mask resulting in the mask image having a multitude of single pixel dots; create an index image wherein each single non-zero pixel of the mask image is replaced by a unique index value beginning with one and increasing by one for each additional non-zero pixel value; use the index image to create a translation array of source image values, one for each index in the index image, in which the value of the translation array is set to the value contained in the source image at a pixel location identical to the pixel location of each index value of the index image; create a waterfall image using the index image as a seed image causing each seed dot to grow into local regions filled with the seed value until the growth encounters an adjacent region from an adjacent seed resulting in the expansion of the plurality of dots in the index image into a plurality of uniformly filled regions forming a Voronoi-type diagram in which each region is filled with the index value from a seed dot contained within that region; create a local-maximum image by replacing each index value in the waterfall image with its corresponding value from the translation array resulting in a Voronoi-type diagram in which each region is uniformly filled with the maximum value of the source image for each region; create a local threshold value in each of the plurality of regions by multiplying the local-maximum image by a threshold parameter with a range of [0.0 . . . 1.0] to create a local threshold value in each image region; create a segmentation image wherein values are zero except where values in the source image region are larger than the values in the local-threshold image to isolate each local region into a single area which contains a maximum value of the region and any adjacent pixels which are above the local maximum value; create a contour around the edge of each segmented image area; create lists of outlines, wherein each outline has its own outline list of points and shape parameters, wherein each outline indicates the presence of a cell; and count the number of cells.

Dependencies are inferred from the claim text.