Cognitive Model of Human Visual Search with Saliency and Scene Context for Real-world Images

Abstract

The previous Adaptive Control of Thought-Rational (ACT-R) cognitive architecture model has limitations in the sense that it cannot accurately predict human visual search for real-world images because scene context which could be as important as saliency is not included. Thus, this study proposed ACT-R cognitive modeling with saliency and scene context in parallel for human visual search. Then, the validation of the model was performed by comparing with eye-tracking experimental data. Results show that the model data was quite well fit with the eye-tracking data. In conclusion, the modeling method proposed in this study should be used, in order to predict actual human visual search using both strategies selectively for real-world image.