Object Detection and Recognition Using Local Quadrant Pattern

Abstract

Object detection and recognition is one of the important techniques in computer vision for searching and scanning and identifying an object in images or videos. Object detection and recognition enters into many important fields where one of the uses of object detection and recognition is to detect region of injury and determine the type of injury. This paper suggested a new effective method called Local Quadrant Pattern (LQP). The proposed method uses a window and passes it on all pixels of the image and uses the pixel direction to arrange the adjacent pixels. It also usesfour code values to encode and then produce a texture feature matrix which is used to detect objects as well as extract features based on magnitude of pixels for image classification. The experiments were conducted on the infected regions in the skin and the results showed the ability of the method to detect regions of infection as well as the high accuracy in the classification of those regions.

Acute lymphoblastic leukemia image segmentation based on modified HSV model

“…Also, [20] used the K-mean clustering, but this time with the Zack algorithm; unfortunately, the authors used only seven images to evaluate the proposed method and got an overall segmentation accuracy of 96.60. Similarly, [21][22][23][24][25] proposed an ALL segmentation approach based on hybrid histogram-based soft covering rough k-means clustering. In addition, they used several features extracted from 350 images derived from the Gray Level Cooccurrence matrix and got an overall segmentation accuracy above 90%.…”