Culture Analysis

Application Type	Segmentation; Counting
Image Characteristics	Monochrome; Good Contrast; Some Speckle Noise
Image Processing Tool(s)	Filtering; Thresholding; Advanced Morphology

The culture analysis example distinguishes the boundaries of each cell in a cluster of cells, separates the cluster into its individual cells, and overlays the cell boundaries over the original image.

Image Buffer: Add Copy (1)—Stores a copy of the original image into Buffer #1 of the image buffer for later use. The final results are overlaid on this image.

Filters: Smoothing—Median—Smoothes the image. The median filter removes any speckle noise that is present in the image. This prevents speckle noise from being highlighted as a detail in the next step.

Filters: Convolution—High Level Detail—Highlights details in the image. The convolution filter highlights regions in the image where there are sharp changes in pixel values. These regions correspond to the boundaries of the cells and other noisy pixels that may be present in the image.

Threshold: Auto-Threshold—Metric—Separates dark regions from the rest of the image. In this case, the dark regions in the image correspond to the cells. The regions corresponding to the medium in which the cells are present appear as background in the foreground binary image.

Adv. Morphology: Remove Small Particles—Removes noise particles in the binary image. Any particle removed by two iterations of the erosion operation is considered to be a noise particle and removed from the image. The rest of the particles are left untouched.

Adv. Morphology: Separate Objects—Separates touching particles in the image. This step helps to separate cells that may be touching each other at a few pixels along their boundaries. At the end of this step, the binary image contains particles corresponding to the cells that need to be counted.

Adv. Morphology: Label Objects—Labels each particle with a unique ID. When you display a labeled image with a binary palette, each particle appears as with a different color. Labeling helps you determine if the cells have been separated correctly.

Basic Morphology: Gradient Out—Finds the outer boundary of each particle in the image.

Operators: Multiply—Multiplies the image by 255 to extend the dynamic of the image from (0,1) to (0,255). Grayscale 255 is represented as white, which becomes the overlap color in the next step.

Operators: Add—Adds the processed image to the original image, which was stored in Buffer #1 of the image buffer. The addition process adds the particle boundary information to the original image, outlining the boundary of each cell.