Shi-Thomasi is a corner detection algorithm improved from Harris' corner detector. The improvement is the way how a certain region within the image is scored (and thus treated as corner or not). Where the Harris-corner detector determines the score $R$ with the eigenvalues $\lambda_1$ and $\lambda_2$ of two regions (the second region is a shifted version of the first one and used to compare to the first one to see if the difference between the two is big enough to say if there is a corner or not) in the following way:

$$R=\det(\lambda_1\lambda_2)-k(\lambda_1+\lambda_2)^2$$

Shi and Thomas just use the minimum of both eigenvalues:

$$R=\min(\lambda_1,\lambda_2)$$

See also:

• http://www.aishack.in/2010/05/the-shi-tomasi-corner-detector/
• @Wikipedia
• Harris and Stephens about their algorithm (pdf)