情報処理学会 第75回全国大会講演要旨
1D-5
Learning-Based Adaptive Algorithm Selection Scheme for Real-World Image Processing
In this paper we propose a real-world information processing universal platform. The main target of the proposed platform is real-world image and information understanding. The platform uses the algorithm selection to chose the best algorithm on a case by case basis. The selection mechanism is obtained using machine learning from a sample data set. We demonstrate advantages of this platform compared to classical single algorithm approach on examples.
