Image Pattern RecognitionUtgave 1

This book describes various types of image patterns for image retrieval. All these patterns are texture dependent. Few image patterns such as Improved directional local extrema patterns, Local Quantized Extrema Patterns, Local Color Oppugnant Quantized Extrema Patterns and Local Mesh quantized extrema patterns are presented. Inter-relationships among the pixels of an image are used for feature extraction. In contrast to the existing patterns these patterns focus on local neighborhood of pixels to creates the feature vector. Evaluation metrics such as precision and recall are calculated after testing with standard databases i.e., Corel-1k, Corel-5k and MIT VisTex database. This book serves as a practical guide for students and researchers.

-The text introduces two models of Directional local extrema patterns viz.,

• Integration of color and directional local extrema patterns
• Integration of Gabor features and directional local extrema patterns.

-Provides a framework to extract the features using quantization method

-Discusses the local quantized extrema collected from two oppugnant color planes

-Illustrates the mesh structure with the pixels at alternate positions.