Thresholding

Thresholding describes the operation of setting values below a given threshold to zero. This may concern all pixels in an image, or amplitudes in a digital signal. Sometimes, the term implies that values above the threshold are set to one, creating a binary image or signal.

Thresholding is often applied to suppress noise, in situations where the signal-to-noise ratio is large. If a high fraction of channels contains only low-amplitude noise, thresholding produces sparse information and may be a powerful step towards data compression. Thresholding with some very simple encoding scheme, like transmitting the sparse channels along with their channel number, is often referred to as zero suppression .