Automatic Image Threshold Segmentation Using Otsu's Method with 2D Attribute Histogram

MATLAB 440K 205 views 0 downloads 1 credits
Tags:

Resource Overview

This paper presents the concept of 2D attribute histogram and proposes an automatic image thresholding method based on 2D attribute histogram using Otsu's algorithm. The implementation involves calculating pixel intensity relationships within local neighborhoods and optimizing inter-class variance through statistical analysis of bidimensional histogram distributions.

Detailed Documentation

In this paper, I propose a novel approach for automatic image threshold segmentation using Otsu's method with a 2D attribute histogram. By introducing the concept of 2D attribute histogram, which simultaneously considers pixel intensity and local attribute information (such as neighborhood mean values), I develop an enhanced automatic thresholding technique. The algorithm implementation typically involves constructing a 2D histogram where one dimension represents pixel intensity and the other captures local contextual features, followed by applying Otsu's maximization of inter-class variance criterion to determine the optimal threshold in the bidimensional space. This method provides significant advantages for academic papers by offering detailed insights into image processing workflows and presenting an innovative methodology that improves segmentation accuracy compared to conventional 1D histogram approaches.

You May Also Like