Start Date
9-5-2019 10:30 AM
End Date
9-5-2019 11:45 AM
Document Type
Event
Description
This paper describes a computational approach to object classification. This paper introduces the reasoning behind pursuing integrals as a mathematical foundation for object detection. In addition this paper outlines the potential steps in an algorithm that can detect objects. The key advantage of this algorithm is its computational complexity. The algorithm can be simplified to a series of subtractions which are O(1) operations versus the much more computationally complex convolutional neural network approach typically applied to classify objects.
DOI
https://doi.org/10.5038/PDPL3817
A Computational Approach to Object Classification
This paper describes a computational approach to object classification. This paper introduces the reasoning behind pursuing integrals as a mathematical foundation for object detection. In addition this paper outlines the potential steps in an algorithm that can detect objects. The key advantage of this algorithm is its computational complexity. The algorithm can be simplified to a series of subtractions which are O(1) operations versus the much more computationally complex convolutional neural network approach typically applied to classify objects.