> For the complete documentation index, see [llms.txt](https://un2c.gitbook.io/un2c-wp/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://un2c.gitbook.io/un2c-wp/un2c-tech/recognition-software-integrations.md).

# Recognition-Software integrations

## <mark style="color:purple;">IMAGE SEGMENTATION</mark>

The acquired image tends to contain unwanted features along with the relevant ones. To eliminate them, a threshold based on the variance of each pixel in the image is performed. Pixels whose intensity (gray level value) is higher than the threshold are taken into account while pixels whose intensity is lower than the threshold are eliminated.

<figure><img src="/files/zHTXLSgRDsIxyFTWNtSP" alt=""><figcaption></figcaption></figure>

## <mark style="color:purple;">IMAGE NORMALIZATION</mark>

Each pixel in the image has a different average-variance ratio. In order to obtain a uniform pattern, a normalization is performed so that the pixels of the image fall within the desired range of gray values.

<figure><img src="/files/QLHiY0Kh3V4z4AYT4S1C" alt=""><figcaption></figcaption></figure>

## <mark style="color:purple;">IMAGE ORIENTATION</mark>

This defines the formation of the image according to the orientation of the peaks for each point. It is done by calculating the gradient of each pixel in the x and y directions, then calculating the orientation by determining the average of the vectors which are orthogonal to the gradient.

<figure><img src="/files/dmUpdzChpdbHzJCLOjAs" alt=""><figcaption></figcaption></figure>

## <mark style="color:purple;">CONSTRUCTION OF THE FREQUENCY IMAGE</mark>

This is done to determine the local frequency (rate of occurrence) of peaks. To do so, we project the gray values of each pixel in the direction perpendicular to the orientation of the peak, and then calculate the number of pixels between consecutive minimums of the waveform (the peaks). Another method is to use the Fourier transform technique.

<figure><img src="/files/XyhWFWdAFcvQK88YU4lv" alt=""><figcaption></figcaption></figure>

## <mark style="color:purple;">IMAGE FILTRATION</mark>

This is done to eliminate unwanted noise. It is performed using a Gabor filter or a Butterworth filter. A basic method is to convolve the image with the filter.

<figure><img src="/files/zsq67GDZufNZekzshgDJ" alt=""><figcaption></figcaption></figure>

## <mark style="color:purple;">IMAGE BINARIZATION</mark>

The filtered image is then converted into a binary image using the thresholding technique, to improve the contrast. It is based on a global threshold, i.e. the pixel value above the threshold is set to 1 and the pixel value below is set to 0.

<figure><img src="/files/uJNjN861ehTH6TLJGldL" alt=""><figcaption></figcaption></figure>

## <mark style="color:purple;">IMAGE THINNING</mark>

This is done to eliminate unwanted noise. It is performed using a Gabor filter or a Butterworth filter. A basic method is to convolve the image with the filter.

<figure><img src="/files/SHvOrul5EG9NrFB44rVW" alt=""><figcaption></figcaption></figure>

Note\*: All recognition system seen above integrated & adapted to suit the digital Anti-counterfeit purposes.&#x20;

Visual exemples provided are to give you a better idea on those individual integrations.