Object recognition and tracking using a classifier comprising cascaded stages of multiple decision trees

作者: Michael A. Pusateri , Maxim Smirnov

DOI:

关键词:

摘要: An image processor comprises first and second hardware accelerators is configured to implement a classifier. The classifier in some embodiments cascaded having plurality of stages with each such stage implementing decision trees. At least one the generate an integral based on given input image, accelerator process patches through or more trees implemented by processor. By way example, illustratively comprise respective front-end back-end processor, calculator accelerator.

参考文章(12)
Brucek Kurdo Khailany, William J. Dally, Thomas Hastings Greer, Carl Thomas Gray, John W. Poulton, Ground-referenced single-ended system-on-package ,(2013)
William Herz, Shivashankar Gurumurthy, Michael L. Schmit, Displayed image improvement ,(2012)
Mihnea Gangea, Peter Corcoran, Leendert Blonk, Illumination detection using classifier chains ,(2008)
Deok-Yeon Kim, Joon-Young Kwak, Byoung Chul Ko, Jae-Yeal Nam, Human Detection Using Wavelet-Based CS-LBP and a Cascade of Random Forests international conference on multimedia and expo. pp. 362- 367 ,(2012) , 10.1109/ICME.2012.124
Neil Birkbeck, Michal Sofka, Shaohua Kevin Zhou, Method and system for evaluation using probabilistic boosting trees ,(2011)