Biochips – fundamentals and applications - Electrochemical Sensors, Biosensors and Their Biomedical Applications - page 324

TABLE 11.4
Software for different segmentation methods

Fixed circle ScanAlyze, GenePix, QuantArray

Adaptive circle GenePix, Dapple

Adaptive shape Spot, region growing, and watershed

Histogram ImaGene, QuantArraym DeArray, and adaptive thresholding

The procedure of processing the images of the microarray consists of addressing,
segmentation, and information extraction.

1. Addressing
Since the microarray spot images have many different sizes and shapes and some of
the spots may not be located at the central position, addressing is essential to fi nd the
centre of the spots. In order to ensure accuracy of the measurements, an automatic
spot detector is used to calculate the spacing between rows and columns of spots, the
overall position of the array in the image and the spot size so that coordinates can be
assigned to each of the spots in the array.

2. Segmentation
Segmentation is the process for classifying the image pixels as either foreground or
background. The fl uorescence intensities are calculated for each spot as a measure of
transcript abundance. A bright spot occurs due to the suffi cient hybridization of the
probe material. The detector is able to detect the fl uorescence of the bound labeled
probe material and record the fl uorescence intensities. The regions occupied by the
bright spot are the foreground. In regions where no hybridization occurs, the intensity is
equal to or less than the background values. Proposed methods for segmentation include
fi xed circle, adaptive circle, adaptive shape, and histogram segmentation. The various
software used to analyze the different segmentation methods are shown in Table 11.4.
In fi xed circle segmentation method, a circle of constant diameter is drawn on all the
spots in the microarray images (Fig. 11.19, see Plate 9 for color version). This method
is easy to implement; however, the fi xed circle segmentation is ineffi cient as spots of
same size and shape rarely occur.
In reality, spots are of different shapes and size. In adaptive circle segmentation,
different circle diameters are assigned to individual spot diameters in the microarray
as can be seen in Fig. 11.20 (see Plate 10 for color version). By doing this, the circle
would fi t the spot perfectly, regardless of the size of the spot. However, the tradeoff is
that much time is required to estimate and compute each of the thousands of spots in
the microarray.
Dapple is the image analysis software used for adaptive circle segmentation. It per-
forms the analysis by estimating the circle diameter for each spot. By taking the sec-
ond derivative of the pixel intensity vs coordinate graph, the edges of the spots can be