• Muthu Poochendu Ratha M
• T.Saravana Kumar

Superpixel segmentation partitions a picture into perceptually coherent segments of Comparable size, namely, superpixels. It is becoming a basic pre-processing step for varies computer vision tasks because of superpixels considerably reduce the number of inputs and provide a meaningful representation for feature extraction. A Proposed pixel-related Gaussian Mixture Model (GMM) to sections pictures into superpixels. GMM could be a weighted sum of Gaussian functions, each one corresponding to a superpixel, to explain the density of every pixel depicted by a random variable. Completely varied from previously proposed GMMs, the weights are constant, and Gaussian functions within the sums are subsets of all the Gaussian functions, resulting in segments of comparable size and an algorithm of linear complexity with respect to the amount of pixels. Additionally to the linear complexity, GMM algorithm is inherently parallel and permits quick execution on multicore systems. Throughout the expectation-maximization iterations of estimating the unknown parameters within the Gaussian functions, Its tends to impose two lower bounds to truncate the eigen values of the covariance matrices, which enables the proposed algorithm to manage the regularity of superpixels.

Superpixel , Image Segmentation , parallel algorithms , Gaussian mixture model , exceptation - maximization

"Object Boundary Detection and Segmentation using Super Pixel Based Gaussian Mixture Model", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 5, page no.388-394, May-2019, Available :http://www.jetir.org/papers/JETIR1905459.pdf

"Object Boundary Detection and Segmentation using Super Pixel Based Gaussian Mixture Model", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 5, page no. pp388-394, May-2019, Available at : http://www.jetir.org/papers/JETIR1905459.pdf