-

Hippocampus is an important structure of the human brain limbic system. The variations in the volume and architecture of this structure have been related to certain neurological diseases.
Accurate and reproducible volume try of hippocampus from MRI is required for diagnosis and treatment of patients. This can be done using the following three steps: 1) determining hippocampus approximate location 2) determining its accurate boundaries and 3) calculating its volume. This paper presents an automated method for the first step. The proposed approach employs a search-based vision for information extraction from MRI and an expert system for information analysis.
In the development of the expert system a shell of expert systems named VP¬EXPERT is used. Information analysis is done using rules that are based on anatomy and symmetry of the human brain. Results show that the proposed approach is promising for object localization from complex medical images. The approach
has been applied to 128 images from 6 patients. In these studies, all of the images without hippocampus were correctly identified. In addition, in 58.3% of the images containing hippocampus, the structure was properly localized.