The role of the image mask

Transfer: https://blog.csdn.net/jacke121/article/details/56668269

With the selected image, or a graphic object, the image treated (wholly or partially) blocking process performed to control the area or the image processing process. For covering a particular image or object referred to as a mask or template. Optical image processing, the mask film may be sufficient, filters and the like. Digital image processing, the mask is a two-dimensional matrix array, and sometimes multi-valued image. Digital image processing, the image mask is mainly used: ① extracting a region of interest, multiplying the image to be processed with a mask created in advance a region of interest, the region of interest to obtain an image, the image values ​​remain constant region of interest, the outer area of ​​the image values ​​are 0. ② shielding effect, with the mask shielding some areas of the image as it does not participate or not participate in the processing of the calculation processing parameters, or only for the shielded area or statistical processing. ③ structural feature extraction, image detection and extraction of the mask with structural features similar to a variable or similarity image matching method. ④ produce special shape of the image.

1, mask defined

       The mask is a binary image consisting of 0 and 1. When applying a mask function, a value of the treated area, the shielded area value 0 is not included in the calculation. By specifying a data value, the data range, finite or infinite value, the region of interest, and annotation files to define the image mask, may be applied to any combination of the above options as input to create a mask. (From Baidu Encyclopedia http://baike.baidu.com/view/2062299.htm#2 )

2, an example of a mask image

                                            

                                                                                              FIG mask 1 (mask) image

 

This intermediate image is focused on a mask 1, there are 87 393 1, i.e. the Fourier domain which can be used to collect data about 1/3 of the original image (512 * 512 = 262144).

3, Baidu library as well as a document on a mask easier to understand

http://wenku.baidu.com/view/bc1407d6b14e852458fb57cd.html

       With the selected image, or a graphic object, the image treated (wholly or partially) blocking process performed to control the area or the image processing process. For covering a particular image or object referred to as a mask or template. Optical image processing, the mask film may be sufficient, filters and the like. Digital image processing, the mask is a two-dimensional matrix array, and sometimes multi-valued image.

       Digital image processing, the image mask is mainly used:

① extracted region of interest, the region of interest with a previously prepared mask image to be processed multiplied by the area of ​​the image of interest, the region of interest image values ​​remain constant, while the outer areas of the image values ​​are 0.

② shielding effect, with the mask shielding some areas of the image as it does not participate or not participate in the processing of the calculation processing parameters, or only for the shielded area or statistical processing.

③ structural feature extraction, image detection and extraction of the mask with structural features similar to a variable or similarity image matching method.

④ produce special shape of the image. With the selected image, or a graphic object, the image treated (wholly or partially) blocking process performed to control the area or the image processing process. For covering a particular image or object referred to as a mask or template.

4, then we take a look at an image (Fourier domain) and mask the result of multiplying (experimental)

                                                

                                                                                              FIG original (lena512 * 512) image

 

                                              

                                                                                              Fourier domain image of FIG. 3 lena

 

                                              

                                                                                              Multiplying the Fourier domain image after the mask of FIG. 4

 

                                              

                                                                                              Inverse Fourier transform image of FIG. 5 back

 

        从图2和图5,我们可以分辨出来掩膜滤掉了部分信息,只选取了掩膜“感兴趣”的区域。

Guess you like

Origin blog.csdn.net/qq_30263737/article/details/90713103