Download - 2010/1/14 陳冠宇
![Page 1: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/1.jpg)
2010/1/14陳冠宇
An Efficient Parallel Eye Detection Algorithm on Facial Color Images∗
![Page 2: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/2.jpg)
Outline
Introduction Idea of Algorithm Proposed Parallel Algorithm Experimental Result Conclusion
![Page 3: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/3.jpg)
IntroductionImage processing applications can be computationally intensive due to large amount of data which is processed and complexity of image processing algorithms.
The best known approach to overcome this issue is using parallel computing in image processing applications.
![Page 4: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/4.jpg)
The proposed algorithm has been examined with MPI and its implementation results on CVL andIranian databases showed that parallel approachreduces the time of detection efficiently.
Introduction
![Page 5: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/5.jpg)
Idea of Algorithm
Eye Map is obtained from a facial image that is transformed into Y CbCr color space
We first build two separate eye maps from facial image, EyeMapC from the chrominance Components and EyeMapL from the luminance component.
![Page 6: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/6.jpg)
![Page 7: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/7.jpg)
It is constructed by:
The 1/3 scaling factor is applied to ensure that the
resulted EyeMapC stays within the range of [0 1]
Parallel EyeMapC Construction
![Page 8: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/8.jpg)
![Page 9: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/9.jpg)
![Page 10: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/10.jpg)
EyeMapL Construction
We use grayscale dilation and erosion with a
hemispheric structuring element to construct
eye map from the luma as follows:
![Page 11: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/11.jpg)
![Page 12: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/12.jpg)
This step is illustrated in Fig. 5. Each processor pis responsible for n/p rows of each EyeMap and multiply the corresponding elements.
![Page 13: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/13.jpg)
Experimental ResultIn this section, we show data related to Effectiveness of proposed parallel algorithm on CVL [15] and Iranian Databases.
We have used a Beowulf cluster with 8 nodesand Ethernet 10/100 network infrastructure. The algorithm have been implemented with LAM/MPI ver. 7.1.14.
![Page 14: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/14.jpg)
![Page 15: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/15.jpg)
![Page 16: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/16.jpg)
![Page 17: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/17.jpg)
MPI Efficiency
![Page 18: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/18.jpg)
![Page 19: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/19.jpg)
Conclusion
This algorithm uses a special color space called
Y CbCr. It constructs two maps in parallel and
merge these two maps together to achieve a final
map.
![Page 20: 2010/1/14 陳冠宇](https://reader033.vdocuments.mx/reader033/viewer/2022050713/568159df550346895dc72b82/html5/thumbnails/20.jpg)
Thanks