Vector Quantization for Machine Vision
This addresses efficiency issues for machine vision practitioners, but appears incremental as it applies an existing compression method to known tasks.
The paper tackles the problem of high computational cost in machine vision tasks by operating directly in the compressed domain using Pyramid Vector Quantization (PVQ), resulting in simplified algorithms like SVMs, CNNs, and HOG features.
This paper shows how to reduce the computational cost for a variety of common machine vision tasks by operating directly in the compressed domain, particularly in the context of hardware acceleration. Pyramid Vector Quantization (PVQ) is the compression technique of choice and its properties are exploited to simplify Support Vector Machines (SVM), Convolutional Neural Networks(CNNs), Histogram of Oriented Gradients (HOG) features, interest points matching and other algorithms.