Meta-recognition is a term used in different contexts, but it generally refers to the ability to think about ones own thinking processes. In the field of computer science, meta-recognition can refer to a performance prediction method for recognition algorithms. In this context, meta-recognition involves analyzing the recognition state information and control information of a recognition system. The goal is to improve the accuracy of recognition systems by analyzing the patterns in recognition scores over a gallery of known images.

In summary, meta-recognition is a term that can refer to the ability to think about ones own thinking processes, as well as a performance prediction method for recognition algorithms in computer science.