**End**

In the SuspiciousConstruction algorithm, by default, each segment's *spci* = 0 (the hypothesis is that no misclassified pixels exist in the segment).

(3) Analyze the suspicious degree

As shown in Figure 2, for a segment, the *spci* value can be calculated based on the pixels in *ISLIC*; *hi*'s corresponding pixels can be grouped as *SP* = {*sp*1, *sp*2, ... , *sp*m}, where *spi* is the pixel number belonging to category *i*, and the inconsistency degree of *SP* can be described using the following formula:

$$\text{inconsistency} (SP) = 1 - \frac{\max\_{i=1\dots m} (sp\_i)}{\sum\_{i} sp\_i} \tag{5}$$

Based on this formula, the value of *hypi* can be expressed as follows:

$$hspc\_i = \begin{cases} 0 & \text{inconsistency } < \alpha \\ 1 & \text{inconsistency } \ge \alpha \end{cases} \tag{6}$$

where α is a threshold value (the default is 0.05). When a segment's *hspci* = 0, the segment's corresponding pixels in *ISLIC* all belong to the same category, which indicates that pixels' features are consistent in both *Iimage* (band value) and *ISLIC* (segment label); in this case, the segmen<sup>t</sup> does not need correction by CRF. In contrast, when a segment's *hspci* = 1, the pixels of the segmen<sup>t</sup> in *ISLIC* belong to different categories, but the pixel's color (band value) is comparatively consistent in *Iimage*; this case may be further subdivided into two situations:

A. Some type of classification error exists in the segmen<sup>t</sup> (such as the classification result deformation problem appearing on an object boundary).

B. The classification result is correct, but the segmen<sup>t</sup> crosses a boundary between objects in *ISLIC* (for example, the number of segments assigned to the SLIC algorithm is too small, and some areas are undersegmented).

In either case, we need to be suspicious of the corresponding segmen<sup>t</sup> and attempt to correct mistakes using CRF. Based on Formulas 5 and 6, the algorithm for analyzing *Icls* using *HList* is as follows (SuspiciousEvaluation Algorithm):
