Maximum excess over sub-arrays distance to assess creative thinking

Torrance Test Creative Thinking (TTCT) figural form was developed to measure creativity based on Guilford's theory. This test measures fluency, flexibility, originality, and elaboration scales. The state-of-the-art assessment involves a human rater to judge respondent responses to creative thinking tests. Such a judging process is time-consuming and subjective preferences. We, therefore, propose a novel assessment of creative thinking to address the issues, partially for elaboration scales. Elaboration scale is the ability to detail and deepen initial ideas which increase their complexity and improves presentation. In image analysis, this scale is determined by estimating the number of objects in an image. The Maximum Excess over Sub Array (MESA) distance predicts image density values by estimating low-level features and sharing the advantages of general regression-based methods, while also maintaining location information. The maximum counting error performs all rectangular subregions as a learning objective function. The result statistically confirmed that the proposed method has no significant difference from human rater judging. In other words, the proposed method is reasonable to judge respondent responses to the creative thinking test.