Algorithms for Evidence Accumulation
Takashi Matsuyama, Vincent Shang-Shouq Hwang · Sigma · 1990
As we have described in Chaper 2, the SIGMA image understanding system consists of three cooperating reasoning modules plus the Question and Answer Module (QAM). The Geometric Reasoning Expert (GRE) performs evidence accumulation for spatial reasoning and constructs the interpretation of the scene. Often, GRE generates hypotheses about undiscovered objects and initiates the top- down verification analysis. A hypothesis is passed to the Model Selection Expert (MSE), which reasons about the most likely appearance of the object. Then, the description of the expected appearance is given to the Low-Level Vision Expert (LLVE), which verifies/refutes its existence in the image. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.