Image bilevel thresholding based on stable transition region set

We utilize the linear system theory to establish a theory model of transition region. With the model, we reveal an important property of transition region, namely the gray level distribution symmetry. Utilizing the property, we propose a new thresholding framework based on stable transition region set. The elements of the stable transition region set are equal or close to each other in the average gray level. As an example of the proposed framework, we have shown that the feature transformation based on the multiscale gradient multiplication technology is an effective means of estimating the threshold. We have performed subjective and objective comparisons on both synthetic and real images. The experimental results show the segmentation quality of the proposed approach is superior to three conventional transition region-based thresholding methods.

Graphical abstract (a) A degraded image (left) and a PTR image (right), where the white pixels denote the PTR of the degraded image; (b) SLTRSLTR and SGTRSGTR, where SLTRSLTR denotes a set of LTRs, and SGTRSGTR is a set of GTRs. The LTRs usually include too many pixels that lie out the PTR (left), or the LTRs only include a few pixels that locate in the PTR (right). On the other hand, the GTRs usually approximate the PTR in the average gray level to some extent (middle). As a result, (c) the L1L1-STS has poor stability (left and right), whereas the L2L2-STS has relatively high stability (middle). The distinction between the L1L1-STS and the L2L2-STS in stability may be used for distinguishing SGTRSGTR from SLTRSLTR.Figure optionsDownload full-size imageDownload as PowerPoint slideHighlights► This research reveals the gray level distribution symmetry of the transition region. ► It proposes a new thresholding method based on stable transition region set. ► New method does not require a fixed empirical parameter to extract transition region. ► New method is independent of the image histogram.