Suitable or efficient 'Face Recognition' method for both low and high resolution Image

2014 8 28
1 个回答
回答已采纳
更新时间：2016 4 14
9 次查看（30 天）

0 个投票

I am working on my fourth year project on Face Detection and Recognition. I am finding it difficult to find a good face recognition method for Low resolution Image as well as high resolution image. I earnestly request you to guide me to find a suitable method and if possible it's algorithm, so that I can conduct my project and submit it in time. Thank you in advance.

1 个评论
显示 -1更早的评论 隐藏 -1更早的评论

tag omda
tag omda 2016-4-14
hi, i'm working on the same project now, did you achieved any progress in that project. if you did it well, could you plz help me?

请先登录，再进行评论。

请先登录，再回答此问题。

 采纳的回答

Image Analyst
Image Analyst 2014-8-28

1 个投票

For the method, I'd suggest you snap a bunch of pictures of people you know to form a "database" - a collection of images in one folder. Then snap a second set of pictures of those people plus some other people. Run your algorithm on those and see if it correctly identifies the people it should and correctly states "not in database" for those people who are not in the database. Then construct an ROC curve. http://en.wikipedia.org/wiki/ROC_curve
For the algorithm, see what the Computer Vision System Toolbox can do, or else go here http://www.visionbib.com/bibliography/contentspeople.html#Face%20Recognition,%20Detection,%20Tracking,%20Gesture%20Recognition,%20Fingerprints,%20Biometrics for published algorithms. If the resolution is low enough, all algorithms will fail. For example you can't identify someone if all you have is a 16 by 16 photo of their face. We don't really help with algorithms unless it's already built-in to a toolbox as a function.

2 个评论
显示 无 隐藏 无

Fahad Mahdi
Fahad Mahdi 2014-8-28
Thank you Sir for your suggestion and guidance. I have tested with my own database (high quality images 1536x2048x3)using PCA algorithm and it works fine. But when I tested the same algorithm with a database with low resolution images (250*250*3), it doesn't work. I mean it can't recognize faces correctly. High Image Database contains images sized around 1200 kb to 1300 kb whereas Low Image Database contains images sized 8 kb to 12 kb. Kindly help me with your valuable suggestion and advice. Thank you.
Image Analyst
Image Analyst 2014-8-28
You would also have to have run the database through the algorithm at the same resolution to see how it behaves with those low res images. There are some algorithms that are fairly scale independent (like SIFT). Not sure what your algorithm does.

请先登录，再进行评论。

更多回答（0 个）

类别

在 帮助中心 和 File Exchange 中查找有关 Semantic Segmentation 的更多信息

产品

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by