Method

BAdaCost with trained on 48x84 images with LDCF features [BdCost48LDCF]
https://github.com/jmbuena/toolbox.badacost.public

Submitted on 13 Jan. 2018 16:38 by
Jose M. Buenaposada (Universidad Rey Juan Carlos)

Running time:0.5 s
Environment:8 cores @ 3.5 Ghz (Matlab + C/C++)

Method Description:
BAdaCost (Multiclass Cost-sensitive Boosting) with Locally Decorrelated Channel Features (LDCF) trained in 48x84 pixels resized images.
Parameters:
See code for details.
Latex Bibtex:
@article{FernandezBaldera2018,
title = "BAdaCost: Multi-class Boosting with Costs ",
journal = "Pattern Recognition ",
volume = "",
number = "",
pages = " - ",
year = "2018",
note = "",
issn = "0031-3203",
doi = "https://doi.org/10.1016/j.patcog.2018.02.022",
url = "https://www.sciencedirect.com/science/article/pii/S0031320318300748",
author = "Antonio Fernández-Baldera and José M. Buenaposada and Luis Baumela",
keywords = "Boosting",
keywords = "Multi-class classification",
keywords = "Cost-sensitive classification",
keywords = "Multi-view object detection ",
abstract = "Abstract We present BAdaCost, a multi-class cost-sensitive classification algorithm. It combines a set of cost-sensitive multi-class weak learners to obtain a strong classification rule within the Boosting framework. To derive the algorithm we introduce CMEL, a Cost-sensitive Multi-class Exponential Loss that generalizes the losses optimized in various classification algorithms such as AdaBoost, SAMME, Cost-sensitive AdaBoost and PIBoost. Hence unifying them under a common theoretical framework. In the experiments performed we prove that \{BAdaCost\} achieves significant gains in performance when compared to previous multi-class cost-sensitive approaches. The advantages of the proposed algorithm in asymmetric multi-class classification are also evaluated in practical multi-view face and car detection problems. "
}

Detailed Results

Object detection and orientation estimation results. Results for object detection are given in terms of average precision (AP) and results for joint object detection and orientation estimation are provided in terms of average orientation similarity (AOS).


Benchmark Easy Moderate Hard
Car (Detection) 77.93 % 67.08 % 51.15 %
Car (Orientation) 77.10 % 66.01 % 50.35 %
This table as LaTeX


2D object detection results.
This figure as: png eps pdf txt gnuplot



Orientation estimation results.
This figure as: png eps pdf txt gnuplot




eXTReMe Tracker