- 1、本文档共24页,可阅读全部内容。
- 2、原创力文档(book118)网站文档一经付费(服务费),不意味着购买了该文档的版权,仅供个人/单位学习、研究之用,不得用于商业用途,未经授权,严禁复制、发行、汇编、翻译或者网络传播等,侵权必究。
- 3、本站所有内容均由合作方或网友上传,本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺!文档内容仅供研究参考,付费前请自行鉴别。如您付费,意味着您自己接受本站规则且自行承担风险,本站不退款、不进行额外附加服务;查看《如何避免下载的几个坑》。如果您已付费下载过本站文档,您可以点击 这里二次下载。
- 4、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等,请点击“版权申诉”(推荐),也可以打举报电话:400-050-0827(电话支持时间:9:00-18:30)。
查看更多
Statistical Color Models with Application to Skin Detection with EM Training
Statistical Color Models with Application to Skin
Detection with EM Training
Anthony Dotterer
3-1-2006
Introduction
Jones and Rehg develop two Gaussian mixture models (GMM) representing color
of skin and not-skin in [1]. However, these models are quite generic; since they were
trained using many different images from the web under arbitrary illumination conditions.
This paper attempts to specialize these models to a single person with a constant
background and constant illumination. To achieve this goal, Expectation Maximization
(EM) is used on a set of training images to fine tune the Gaussian mixture models.
Setup
In order to perform EM, a Gaussian mixture model must be initialized. The
GMMs derived in [1] are used from the code provided by Dr. Collins. There are 32
models, where 16 are skin matching models and the other 16 are non-skin, background
models. The following figures show some input, the skin mask output, and the image
mask output.
Illustration 1: Input picture 1
Illustration 2: Skin mask of picture 1
Illustration 3: Image masked for picture 1
The following table contains the parameters of the GMMs in [1].
Model Red
Mean
Green
Mean
Blue
Mean
Red
Variance
Green
Variance
Blue
Variance
Weight
1 73.53 29.94 17.76 765.40 121.44 112.80 0.0294
2 249.71 233.94 217.49 39.94 154.44 396.05 0.0331
3 161.68 116.25 96.95 291.03 60.48 162.85 0.0654
4 186.07 136.62 114.40 274.95 64.60 198.27 0.0756
5 189.26 98.37 51.18 633.18 222.40 250.69 0.0554
6 247.00 152.20 90.84 65.23 691.53 609.92 0.0314
7 150.10 72.66 37.76 408.63 200.77 257.57 0.0454
8 206.85 171.09 156.34 530.08 155.08 572.79 0.0469
9 212.78 152.82 120.04 160.57 84.52 243.90 0.0956
10 234.87 175.43 138.94 163.80 121.57 279.22 0.0763
11 151.19 97.74 74.59 425.40 73.56 175.11 0.1100
12 120.52 77.55 59.82 330.45 70.34 151.82 0.0676
13 192.20 119.62 82.32 152.76 92.14 259.15 0.0755
14 214.29 136.08 87.24 204.90 140.17 270.19 0.0500
15 99.57 54.33 38.06 448.13 90.18 151.29 0.0667
16 238.88 2
您可能关注的文档
- Related-Key and Slide Attacks Analysis, Connections, and Improvements Extended Abstract.pdf
- Relative Quantifications Data Management& Analysis Settings.pdf
- Relativistic Instant--Form Approach to the Structure of Two-Body Composite Systems. II. Non.pdf
- Relativistic stars in f(R) gravity, and absence thereof.pdf
- reinventing_your_business_model.pdf
- Relief Mapping of Non-Height-Field Surface Details.pdf
- Renormalization of number density in nonequilibrium quantum-field theory and absence of pin.pdf
- RENK 齿轮箱产品介绍.pdf
- Report on an all-sky LIGO search for periodic gravitational waves in the S4 data.pdf
- Reliable Multi-Path Routing Schemes for Real-Time Streaming.pdf
文档评论(0)