High-Level Synthesis of Online K-Means Clustering Hardware for a Real-Time Image Processing Pipeline

Abstract

The growing need for smart surveillance solutions requires that modern video capturing devices to be equipped with advance features, such as object detection, scene characterization, and event detection, etc. Image segmentation into various connected regions is a vital pre-processing step in these and other advanced computer vision algorithms. Thus, the inclusion of a hardware accelerator for this task in the conventional image processing pipeline inevitably reduces the workload for more advanced operations downstream. Moreover, design entry by using high-level synthesis tools is gaining popularity for the facilitation of system development under a rapid prototyping paradigm. To address these design requirements, we have developed a hardware accelerator for image segmentation, based on an online K-Means algorithm using a Simulink high-level synthesis tool. The developed hardware uses a standard pixel streaming protocol, and it can be readily inserted into any image processing pipeline as an Intellectual Property (IP) core on a Field Programmable Gate Array (FPGA). Furthermore, the proposed design reduces the hardware complexity of the conventional architectures by employing a weighted instead of a moving average to update the clusters. Experimental evidence has also been provided to demonstrate that the proposed weighted average-based approach yields better results than the conventional moving average on test video sequences. The synthesized hardware has been tested in real-time environment to process Full HD video at 26.5 fps, while the estimated dynamic power consumption is less than 90 mW on the Xilinx Zynq-7000 SOC.

A Combined Offline and Online Algorithm for Real-Time and Long-Term Classification of Sheep Behaviour: Novel Approach for Precision Livestock Farming

“…Several different methods exist for such online learning, including incremental support vector machines, online random forest, incremental vector quantization and stochastic gradient descent, among others. In this study, we used an unsupervised k-means online clustering algorithm [29]. The unsupervised k-means online clustering algorithm was implemented in Matlab [30] using the k-means built in function.…”

Optical Signature Analysis of Liver Ablation Stages Exploiting Spatio-Spectral Imaging

“…The K-means clustering is the simplest unsupervised learning algorithm that is capable to solve the distinguished clustering issues. K-means algorithm is prevalent in rapid decision making approach for its simplicity, adequacy as well as being moderate although it had a steady presentation across various issues [54,62,63] Despite the way that the time inconvenience was direct to the information size, customary k-implies is so far not admirably compelling to manage a web-scale data [61]. The strategy follows a basic as well as a simple approach to characterize a given data set via a specific number of clusters settled an earlier [64].…”

Unsupervised color image segmentation: A case of RGB histogram based K-means clustering initialization

“…According to [39], regions should be uniform, the boundaries of the regions must be simple, not ragged and adjacent regions must have a significant difference according to the considered uniformity criteria. In classical clustering algorithms such as K-means and Fuzzy C-means (FCM), objects are categorized into different classes based on similar attributes of the data objects [40][41][42][43]. The K-means technique is considered one of the simplest methods with a fast convergence [43].…”