Exploiting Data Parallelism in the yConvex Hypergraph Algorithm for Image Representation using GPGPUs

TL;DR

This paper presents a parallel implementation of the yConvex Hypergraph model for image region-of-interest detection using GPGPU, significantly improving performance on high-resolution images.

Contribution

It introduces a CUDA-based parallel approach for the yCHG model, demonstrating substantial speedups on large satellite images compared to serial execution.

Findings

Parallel implementation outperforms serial on high-resolution images (2x-10x speedup)

Performance is unaffected by the number of hyperedges in the ROI

Serial implementation is better for smaller images

Abstract

To define and identify a region-of-interest (ROI) in a digital image, the shape descriptor of the ROI has to be described in terms of its boundary characteristics. To address the generic issues of contour tracking, the yConvex Hypergraph (yCHG) model was proposed by Kanna et al [1]. In this work, we propose a parallel approach to implement the yCHG model by exploiting massively parallel cores of NVIDIA's Compute Unified Device Architecture (CUDA). We perform our experiments on the MODIS satellite image database by NASA, and based on our analysis we observe that the performance of the serial implementation is better on smaller images, but once the threshold is achieved in terms of image resolution, the parallel implementation outperforms its sequential counterpart by 2 to 10 times (2x-10x). We also conclude that an increase in the number of hyperedges in the ROI of a given size does not impact the performance of the overall algorithm.