Abstract
State-of-the-art pedestrian detectors are capable of finding humans in images with reasonable accuracy. However, accurate object detectors such as Integral Channel Features (ICF) do not provide good reliability; they are unable to identify detections which they are less confident (or more uncertain) about. We apply existing methods for generating probabilistic measures from classifier scores (such as Piatt exponential scaling and Isotonic Regression) and compare these to Gaussian Process classifiers (GPCs), which can provide more informative predictive variance. GPCs are less accurate than ICF classifiers, but GPCs and Adaboost with Piatt scaling both provide improved reliability over existing methods.
Original language | English |
---|---|
Title of host publication | 2014 Sensor Signal Processing for Defence |
Publisher | IEEE |
Number of pages | 5 |
ISBN (Print) | 9781479952946 |
DOIs | |
Publication status | Published - 31 Oct 2014 |
Event | 4th Sensor Signal Processing for Defence 2014 - Edinburgh, Edinburgh, United Kingdom Duration: 8 Sept 2014 → 9 Sept 2014 |
Conference
Conference | 4th Sensor Signal Processing for Defence 2014 |
---|---|
Abbreviated title | SSPD 2014 |
Country/Territory | United Kingdom |
City | Edinburgh |
Period | 8/09/14 → 9/09/14 |
ASJC Scopus subject areas
- Signal Processing
- Computer Science Applications
- Electrical and Electronic Engineering