扩展的旋转哈尔特征在快速目标检测中的应用
需积分: 10 154 浏览量
更新于2024-09-08
收藏 268KB PDF 举报
"哈尔特征 英文版 - An Extended Set of Haar-like Features for Rapid Object Detection"
哈尔特征是一种在计算机视觉领域广泛使用的图像处理技术，特别适用于物体检测，尤其是人脸识别。这些特征源于数学家瓦尔德（Wald）的哈尔函数，它们通过分析图像中像素的灰度差异来捕获图像的局部特性。描述哈尔特征的关键在于它们的构造方式：通过将图像分割成黑（低值）和白（高值）的矩形区域，然后计算这些矩形的像素和之差来形成特征向量。
哈尔特征分为三类：边缘特征、线性特征、中心特征和对角线特征。边缘特征关注图像的边界变化，线性特征捕捉水平或垂直方向的强度变化，而中心和对角线特征则用于描述更复杂的方向变化。例如，在人脸识别中，眼睛区域通常比脸颊暗，鼻梁两侧比鼻梁本身暗，这些都可以通过不同形状和大小的矩形组合的哈尔特征来表示。
Rainer Lienhart和Jochen Maydt在他们的论文"An Extended Set of Haar-like Features for Rapid Object Detection"中进一步扩展了哈尔特征的概念，引入了一组旋转的哈尔特征。这些新的特征不仅丰富了原有的简单特征集，而且能更有效地计算，尤其对于检测那些具有非垂直或水平边缘的对象更有优势。通过使用旋转哈尔特征，他们的人脸检测器在保持相同检测精度的同时，平均降低了10%的误报率。
此外，他们还提出了一种针对已训练好的级联分类器的后优化过程，可以进一步平均降低12.5%的误报率。这种方法对于提升检测算法的性能至关重要，因为它减少了在大量无目标图像上浪费的计算资源，从而提高了整体检测效率。
哈尔特征是计算机视觉中的一个重要工具，尤其在快速目标检测中，如人脸识别、行为识别等领域。通过不断扩展和优化，哈尔特征能够适应更复杂的图像结构，提高检测的准确性和效率。
An Extended Set
of
Haar-like Features
for
Rapid Object Detection
Rainer Lienhart and Jochen
Maydt
Intel Labs, Intel Corporation, Santa Clara, CA
95052,
USA
Rainer. Lienhart@intel.com
ABSTRACT
Recently Viola et al. [5] have introduced a rapid object detection.
scheme based on a boosted cascade of simple feature classifiers. In
this
paper we introduce a novel set of rotated haar-like features.
These novel features significantly enrich the simple features of (51
and can also be calculated efficiently. With these new rotated
features our sample face detector shows
off
on average a
10%
lower
false alarm rate at a given hit rate. We also present a novel post
optimization procedure for
a
given boosted cascade improving on
average the false alarm rate further by 12.5%.
1
Introduction
Recently Viola et al. have proposed a multi-stage classification
procedure that reduces the processing time substantially while
achieving almost the same accuracy
as
compared to a much slower
and more complex single stage classifier (51.
This
paper extends
their rapid object detection framework in two important ways:
Firstly, their basic and over-complete set of haar-like feature is
extended by an efficient set of 45" rotated features, which add
additional domain-knowledge to the leaming framework and which
is otherwise hard to learn. These novel features can be computed
rapidly at
all
scales in constant time. Secondly, we derive a new post-
optimization procedure for a given boosted classifier that improves
its performance significantly.
2
Features
The main purpose of using features instead of raw pixel values
as
the
input to a leaming algorithm is to reducehncrease the in-clasdout-
of-class
variability
compared
10
the
raw
input data, and thus making
classification easier. Features usually encode knowledge about the
domain. which
is
difncult to
learn
from
a
raw finite set of input data.
The complexity of feature evaluation is
also
a very important aspect
since almost all object detection algorithms slide a fixed-size
window at all scales over the input image.
As
we will see. our
features can be computed at any position and any scale in the same
constant time. Only
8
table lookups are needed.
2.1
Feature
Pool
Our feature
pool
was inspired by the over-complete haar-like
features used by Papageorgiou
el
al.
in
14.31
and their very fast
computation scheme proposed by Viola el
al.
in (51, and
is
a
generalization of their work.
Let
us
assume that the basic unit for testing for presence of an object
is a window
of
M*H
pixels.
Also
assume that we have a very fast
way of computing the sum of pixels of any upright and 45" rotated
rectangle inside the window.
A
rectangle is specified by the tuple
r=(x,y,w,h,a) with
05x,x+tdW.
O5y,y+KH.
x,y>O.
w,h>O,
and
(IE
{0",45')
and its pixel sum
Is
denoted by RecSum(r), Two
examples of such rectangles are given in Figure
1.
Our raw feature set
is
then the set of all possible features of the form
featmi o,.RecSUm(r,),
IE
141.
...,NI
where the weights
op
CR
,
the rectangles
r,,
and
N
are arbitrarily
Window
upright rectangle
9.1.
Exampledanu~nand45'r~~reaangle.
chosen.
This raw feature set is (almost) infinitely large. For practical reasons,
it is reduced as follows:
1.
Only weighted combinations of pixel
sums
of
two
rectangles are
considered (i.e..
N=
2
).
2.
The weights have opposite signs, and are used to compensate for
the difference in area size between the two rectangles. Thus, for
non-overlapping rectangles we have
-wo.Area(ro)= wi.Area(rl)
.
Without restrictions we can set
wo=-l
and get wi=Area(roYArea(r,).
3.
The features mimic haar-like features and early features of the
human visual pathway such as center-surround and directional
responses.
These cestrictions lead
us
to
the
14
feature
prototypes shown
in
Figure
2:
Eight line features. and
*
Two center-surround features.
These prototypes are scaled independently in vertical and horizontal
direction in order to generate a rich, over complete set of features.
Note that the line features can be calculated by two rectangles only.
Hereto it is assumed that the first rectangle
ro
encompasses the
black and white rectangle and the second rectangle
rl
represents the
black area.
For
instance, line feature (Za) with total height of 2 and
width of
6
at the top left corner
(5.3)
can be written
as
Four edge features.
fea1ure~-1.RecSum(5,3,6,2,0")+3~RecSum(7,3,2,2,0')
Only features (la), (Ib). (2a). (23 and (43 of Figure
2
have been
used by (3.4.51. In our experiments the additional features
significantly enhanced the expressional power of the learning
system and consequently improved the performance
of
the object
detection system. Feature (4a) was not used since it is well
approximated by feature (29) and @e).
NUMBER
OF
FEATURES.
The number of features derived from each
prototype is quite large and differs from prototype to prototype and
can be calculated as follows. Let X=iWwj and
Y=LWhl
be the
maximum scaling factors in wand ydirection.
A
upright feature of
size
wxh
then generates
0-7803-7622-6/02/$17.00
a2002
IEEE
I
-
900
IEEE
ICIP
2002
下载后可阅读完整内容，剩余3页未读，立即下载
499 浏览量
点击了解资源详情
点击了解资源详情
109 浏览量
144 浏览量
2021-07-08 上传
2021-08-05 上传
2013-08-19 上传
185 浏览量
scorqing
- 粉丝: 0
我的内容管理
展开
最新资源
- 网页自动刷新工具 v1.1 - 自定义时间间隔与关机
- pt-1.4协程源码深度解析
- EP4CE6E22C8芯片三相正弦波发生器设计与实现
- 高效处理超大XML文件的查看工具介绍
- 64K极限挑战：国际程序设计大赛优秀3D作品展
- ENVI软件全面应用教程指南
- 学生档案管理系统设计与开发
- 网络伪书：社区驱动的在线音乐制图平台
- Lettuce 5.0.3中文API文档完整包下载指南
- 雅虎通Yahoo! Messenger v0.8.115即时聊天功能详解
- 将Android手机转变为IP监控摄像机
- PLSQL入门教程：变量声明与程序交互
- 掌握.NET三层架构：实例学习与源码解析
- WPF中Devexpress GridControl分组功能实例分析
- H3Viewer: VS2010专用高效帮助文档查看工具
- STM32CubeMX LED与按键初始化及外部中断处理教程
