Intelligent Shape Feature Extraction and Indexing for Efficient

Introduction Problems with traditional methods of image indexing [4] have led to great interest in Content-Based Image Retrieval (CBIR) techniques, which employ automatically derived features such as color, texture, and shape as search criteria [8]. The CBIR problem is a challenging topic of current research [5] and is of growing interest in medical applications such as medical records systems. The ability to use non-text information to search for images having certain characteristics would add a powerful new dimension to medical imagery. Medical image databases usually contain a wide variety of image modalities: X-ray, magnetic resonance, computed tomography, and ultrasound being the most common. For large image databases, practical CBIR methods must be very fast and efficient in their query and retrieval operations. To help achieve this, our research concentrates upon developing useful features and indexing them for efficient search.