OpenCL-accelerated object classification in video streams using Spatial Pooler of Hierarchical Temporal Memory

TL;DR

This paper presents an OpenCL-accelerated object classification system in video streams using Hierarchical Temporal Memory, achieving high accuracy and significant speed-up on selected computational parts, though overall acceleration is limited by data transfer overhead.

Contribution

The paper introduces an OpenCL implementation for accelerating HTM-based object classification in video streams, demonstrating substantial kernel speed-ups and analyzing performance trade-offs.

Findings

F1 score up to 0.95 achieved.

Kernel speed-up up to 632x on GPU.

Overall acceleration limited to 6.5x due to transfer times.

Abstract

We present a method to classify objects in video streams using a brain-inspired Hierarchical Temporal Memory (HTM) algorithm. Object classification is a challenging task where humans still significantly outperform machine learning algorithms due to their unique capabilities. We have implemented a system which achieves very promising performance in terms of recognition accuracy. Unfortunately, conducting more advanced experiments is very computationally demanding; some of the trials run on a standard CPU may take as long as several days for 960x540 video streams frames. Therefore we have decided to accelerate selected parts of the system using OpenCL. In particular, we seek to determine to what extent porting selected and computationally demanding parts of a core may speed up calculations. The classification accuracy of the system was examined through a series of experiments and the performance was given in terms of F1 score as a function of the number of columns, synapses, $min\_overlap$ and $winners\_set\_size$. The system achieves the highest F1 score of 0.95 and 0.91 for $min\_overlap=4$ and 256 synapses, respectively. We have also conduced a series of experiments with different hardware setups and measured CPU/GPU acceleration. The best kernel speed-up of 632x and 207x was reached for 256 synapses and 1024 columns. However, overall acceleration including transfer time was significantly lower and amounted to 6.5x and 3.2x for the same setup.