Face image preprocesaing filtering

1 次查看（过去 30 天）
AP
AP 2020-4-12
评论： AP 2020-4-15
I want to filter face image.
How to filter image with mathematical calculations?
How to remove blur, noise in face area?
How to denoise image?

请先登录，再回答此问题。

回答（2 个）

KALYAN ACHARJYA
KALYAN ACHARJYA 2020-4-12
You did not attach the sample image. The question is quite broad, it is quit unfair to answer image processing noise related question without observing the noisy image.
Here are some suggestions-
  1. Carefully observe it, if the involved noise is familiar (like impulse noise, salt and pepper, some casse gaussian too), you may predict it directly by observing the images.
  2. If not (Step 1), See the Historam of the image, from the distrbution map of the histogram, you may get clue about the type of noise involved on images, afterwards do estimation of noise type. For better understanding, please refer the Gonzalez Book (Chapters Image Enghancement or Image Restoration) or this PPT directly
  3. If you get the idea, create an approprite h using fspecial
  4. Then proceed for imfilter with h, defined in step 3.
  5. Some cases, you may have check with multiple h (custom also), see the results, where you get the best results.
Good Wishes!
  1 个评论
AP
AP 2020-4-13
Thank you sir for your reply.
Currently I am using database from criminal images. In my case, the images are blurred,low contrast,no complete visibility. I have tried filtering such images with imfilter,fspecial filter but i am not getting proper results. I think i should use some preprocessing along with filtering. i tried histogram but it will not be sufficient for all cases.
I need to figure this out. please suggest me similar methods like histogram for clearing face images from whole image and make the image clear ,visible.
Thank you in advance.

请先登录，再进行评论。

Image Analyst
Image Analyst 2020-4-12
编辑：Image Analyst 2020-4-12
Like Kalyan said, really broad question. We can't give a whole course here, so I'll just point you to some functions:
Deblur: conv2(), imfilter(), deconvlucy()
Denoise: medfilt2(), imfilter(), imnlmfilt() (non local means), conv2(), imbilatfilt(), imdiffusefilt(), etc.
Also see my face pixelating/blurring script in this link
  3 个评论
显示 1更早的评论
Image Analyst
Image Analyst 2020-4-13
Taking the histogram just describes the image - it does not enhance it, though the histogram might be used in deciding some parameters of the enhancement algorithm.
Again, enhancement is also such a broad topic. You could do a simple contrast stretch with imadjust(), or you could do something like adapthisteq(). If you want more algorithms, you could browse the papers here: Vision Bibliography
AP
AP 2020-4-15
Thank you sir for your reply. I am checking all methods you have suggested. I will inform you when i need any help in the near future. Thank you once again.

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Image Filtering 的更多信息

标签

产品

Community Treasure Hunt

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

Start Hunting!

Translated by