Content Based Image Retrieval Using CS-LBP, LPQ and GLCM: Basics, Concept and Novel Algorithm - Softcover

Inhaltsangabe

A new image retrieval technique based on Content Based Image Retrieval has been introduced. This technique is useful for different kind of datasets. CSLBP (center-symmetric local binary patterns) has been extracted from the original image to obtain the local information. LPQ (local phase quantisation) is also insensitive against another image degradation, blur effect. This is common illumination and blurs insensitive feature extraction. Co-occurrence of pixel pairs in local pattern map has been observed in different directions and distances using gray level co-occurance matrix. For similarity measure, four different similarity measure techniques i.e. Chi-square, Manhattan, Canbarra and Euclidian distances are used in which Canbara gives better results than the rest. Out of four, Euclidian distance shows poor performances in comparable to other similarity measures. Experimental results show that Canbarra based bio-medical retrieval system gives about 99-100 % accuracy rates.

Die Inhaltsangabe kann sich auf eine andere Ausgabe dieses Titels beziehen.