I am trying to calculate the absolute difference of two images using a mask so only a region of the images is considered in calculating the difference. But OpenCV does not have the mask part in its function. I saw this question but did not work for me. I am trying to multiply the result in the mask so that only the specified region remains.
code:
Mat region = //a grayscale image containing a region of 255 and the rest is zero
Mat img1, img2 = //two images of the same size as the region image and of type CV_8UC1
Mat mask = region / 255; //to make a binary mask
Mat difference = Mat::zeros(region .rows, region .cols, CV_8UC1);
cv::absdiff(img1, img2, difference);
difference = difference * mask;
if (!difference.empty()) imshow("difference", difference);
When I try this, I get an error.
error:
Error: Assertion failed (a_size.width == len) in cv::gemm
which happens here:
inline
Mat& Mat::operator = (const MatExpr& e)
{
e.op->assign(e, *this);
return *this;
}
difference * mask meaning that you are performing Matrix multiplication, in this case the height of difference must be the same as width of mask, if you want to perform an Element wise multiplication you should call difference.mul(mask)
Related
I am trying to update part of a Mat based on another Mat. For example, I want to select a part of img that is not zero in mask and add a constant value to it. When I try this:
Mat mask = imread("some grayscale image with a white area in a black background", IMREAD_GRAYSCALE);
Mat img = Mat::zeros(mask.rows, mask.cols, CV_8UC1);
Mat bnry, locations;
threshold(mask, bnry, 100, 255, THRESH_BINARY);
findNonZero(bnry, locations);
img(locations) += 5;
I get this error:
Error: Assertion failed ((int)ranges.size() == d)
img and mask have the same size.
How can I select an area of an image based on another image (mask)?
Many of the OpenCV functions will support mask in default, in other word you don't need to find non zero values and based on that doing sum operation, you just need to use cv::add function that in default support using mask as an argument,
cv::add(img,10,img,mask); // 10 is an arbitrary constant value
And about your code
img(locations) += 5;
As far as I know we don't have any like this overloaded operator+ in OpenCV to use.
I am creating a code for change detection in C++ using OpenCV but this code shows runtime error if I change the the image
void MainWindow::on_pushButton_2_clicked()
{
cv::Mat input1 = cv::imread("C:\\Users\\trainee2017233\\Desktop\\pre-post\\sulamani_ms1p1_pre_gref.tif");
cv::Mat input2 = cv::imread("C:\\Users\\trainee2017233\\Desktop\\post-post\\sulamani_ms1p1_pre_gref.tif");
cv::Mat diff;
cv::absdiff(input1, input2, diff);
cv::Mat diff1Channel;
// WARNING: this will weight channels differently! - instead you might want some different metric here. e.g. (R+B+G)/3 or MAX(R,G,B)
cv::cvtColor(diff, diff1Channel, CV_BGR2GRAY);
float threshold = 30; // pixel may differ only up to "threshold" to count as being "similar"
cv::Mat mask = diff1Channel < threshold;
cv::imshow("similar in both images" , mask);
// use similar regions in new image: Use black as background
cv::Mat similarRegions(input1.size(), input1.type(), cv::Scalar::all(0));
// copy masked area
input1.copyTo(similarRegions, mask);
cv::imshow("input1", input1);
cv::imshow("input2", input2);
cv::imshow("similar regions", similarRegions);
cv::imwrite("../outputData/Similar_result.png", similarRegions);
cv::waitKey(0);
}
when I am writing both images as the same image then no error is there but while changing them to different images it shows the error
OpenCV Error: Sizes of input arguments do not match (The operation is neither 'array op array' (where arrays have the same size and the same number of channels), nor 'array op scalar', nor 'scalar op array') in arithm_op, file D:\opencv\sources\modules\core\src\arithm.cpp, line 659
Here input1 and input2 should be of the same size for the function absdiff
...
cv::resize(input2, input2, input1.size());
cv::Mat diff;
...
I am attempting to scale a grayscale image of type 8UC1 by 1.0f/255,
by the following operation
image.convertTo(image,CV_32F,1.0f/255,0); //convert and scale
After inspecting the output of the above, I find that all values are too close to zero. For instance, at a point where the value should be 0.2784, I'm getting 1.23417e-06.
So, I tried to see if I could undo the scaling and get the input back i.e. multiplying the result from above by 255, using
cv::imwrite("undo_scaling.jpg",image*255); //rescale and write to disk
strangely, the input image can be reconstructed.
Where am I going wrong with the scaling operation?
EDIT
The following is my attempt to preprocess an image. That involves the followings steps
Apply a mask
Crop the result
Finally, scale the pixel values by 255
I use the following code:
cv::Mat maskCrop(std::string imageName, std::string maskName)
{
cv::Mat image,mask,final_image;
image = cv::imread( imageName, CV_LOAD_IMAGE_GRAYSCALE);
mask = cv::imread( maskName,CV_LOAD_IMAGE_GRAYSCALE);
cv::resize(image, image, mask.size()); // make the size of mask and image same
cv::bitwise_and(image, mask, final_image); //Apply mask
// define rectangular window for cropping
int offset_x = 1250; // top left corner, for cropping
int offset_y = 1550;
cv::Rect roi;
roi.x = offset_x;
roi.y = offset_y;
roi.width = 550;
roi.height = 650;
// Crop the original image to the defined ROI //
cv::Mat crop = dstImage(roi);
crop.convertTo(crop,CV_32F,1.0f/255,0);
return crop;
}
Below is the input image:
The following is the mask to be applied on it:
I create a Bird-View-Image with the warpPerspective()-function like this:
warpPerspective(frame, result, H, result.size(), CV_WARP_INVERSE_MAP, BORDER_TRANSPARENT);
The result looks very good and also the border is transparent:
Bird-View-Image
Now I want to put this image on top of another image "out". I try doing this with the function warpAffine like this:
warpAffine(result, out, M, out.size(), CV_INTER_LINEAR, BORDER_TRANSPARENT);
I also converted "out" to a four channel image with alpha channel according to a question which was already asked on stackoverflow:
Convert Image
This is the code: cvtColor(out, out, CV_BGR2BGRA);
I expected to see the chessboard but not the gray background. But in fact, my result looks like this:
Result Image
What am I doing wrong? Do I forget something to do? Is there another way to solve my problem? Any help is appreciated :)
Thanks!
Best regards
DamBedEi
I hope there is a better way, but here it is something you could do:
Do warpaffine normally (without the transparency thing)
Find the contour that encloses the image warped
Use this contour for creating a mask (white values inside the image warped, blacks in the borders)
Use this mask for copy the image warped into the other image
Sample code:
// load images
cv::Mat image2 = cv::imread("lena.png");
cv::Mat image = cv::imread("IKnowOpencv.jpg");
cv::resize(image, image, image2.size());
// perform warp perspective
std::vector<cv::Point2f> prev;
prev.push_back(cv::Point2f(-30,-60));
prev.push_back(cv::Point2f(image.cols+50,-50));
prev.push_back(cv::Point2f(image.cols+100,image.rows+50));
prev.push_back(cv::Point2f(-50,image.rows+50 ));
std::vector<cv::Point2f> post;
post.push_back(cv::Point2f(0,0));
post.push_back(cv::Point2f(image.cols-1,0));
post.push_back(cv::Point2f(image.cols-1,image.rows-1));
post.push_back(cv::Point2f(0,image.rows-1));
cv::Mat homography = cv::findHomography(prev, post);
cv::Mat imageWarped;
cv::warpPerspective(image, imageWarped, homography, image.size());
// find external contour and create mask
std::vector<std::vector<cv::Point> > contours;
cv::Mat imageWarpedCloned = imageWarped.clone(); // clone the image because findContours will modify it
cv::cvtColor(imageWarpedCloned, imageWarpedCloned, CV_BGR2GRAY); //only if the image is BGR
cv::findContours (imageWarpedCloned, contours, CV_RETR_EXTERNAL, CV_CHAIN_APPROX_NONE);
// create mask
cv::Mat mask = cv::Mat::zeros(image.size(), CV_8U);
cv::drawContours(mask, contours, 0, cv::Scalar(255), -1);
// copy warped image into image2 using the mask
cv::erode(mask, mask, cv::Mat()); // for avoid artefacts
imageWarped.copyTo(image2, mask); // copy the image using the mask
//show images
cv::imshow("imageWarpedCloned", imageWarpedCloned);
cv::imshow("warped", imageWarped);
cv::imshow("image2", image2);
cv::waitKey();
One of the easiest ways to approach this (not necessarily the most efficient) is to warp the image twice, but set the OpenCV constant boundary value to different values each time (i.e. zero the first time and 255 the second time). These constant values should be chosen towards the minimum and maximum values in the image.
Then it is easy to find a binary mask where the two warp values are close to equal.
More importantly, you can also create a transparency effect through simple algebra like the following:
new_image = np.float32((warp_const_255 - warp_const_0) *
preferred_bkg_img) / 255.0 + np.float32(warp_const_0)
The main reason I prefer this method is that openCV seems to interpolate smoothly down (or up) to the constant value at the image edges. A fully binary mask will pick up these dark or light fringe areas as artifacts. The above method acts more like true transparency and blends properly with the preferred background.
Here's a small test program that warps with transparent "border", then copies the warped image to a solid background.
int main()
{
cv::Mat input = cv::imread("../inputData/Lenna.png");
cv::Mat transparentInput, transparentWarped;
cv::cvtColor(input, transparentInput, CV_BGR2BGRA);
//transparentInput = input.clone();
// create sample transformation mat
cv::Mat M = cv::Mat::eye(2,3, CV_64FC1);
// as a sample, just scale down and translate a little:
M.at<double>(0,0) = 0.3;
M.at<double>(0,2) = 100;
M.at<double>(1,1) = 0.3;
M.at<double>(1,2) = 100;
// warp to same size with transparent border:
cv::warpAffine(transparentInput, transparentWarped, M, transparentInput.size(), CV_INTER_LINEAR, cv::BORDER_TRANSPARENT);
// NOW: merge image with background, here I use the original image as background:
cv::Mat background = input;
// create output buffer with same size as input
cv::Mat outputImage = input.clone();
for(int j=0; j<transparentWarped.rows; ++j)
for(int i=0; i<transparentWarped.cols; ++i)
{
cv::Scalar pixWarped = transparentWarped.at<cv::Vec4b>(j,i);
cv::Scalar pixBackground = background.at<cv::Vec3b>(j,i);
float transparency = pixWarped[3] / 255.0f; // pixel value: 0 (0.0f) = fully transparent, 255 (1.0f) = fully solid
outputImage.at<cv::Vec3b>(j,i)[0] = transparency * pixWarped[0] + (1.0f-transparency)*pixBackground[0];
outputImage.at<cv::Vec3b>(j,i)[1] = transparency * pixWarped[1] + (1.0f-transparency)*pixBackground[1];
outputImage.at<cv::Vec3b>(j,i)[2] = transparency * pixWarped[2] + (1.0f-transparency)*pixBackground[2];
}
cv::imshow("warped", outputImage);
cv::imshow("input", input);
cv::imwrite("../outputData/TransparentWarped.png", outputImage);
cv::waitKey(0);
return 0;
}
I use this as input:
and get this output:
which looks like ALPHA channel isn't set to ZERO by warpAffine but to something like 205...
But in general this is the way I would do it (unoptimized)
I want to apply a binary mask to a color image.
Please provide a basic code example with proper explanation of how the code works.
Also, is there some option to apply a mask permanently so all functions operate only within the mask?
While #perrejba s answer is correct, it uses the legacy C-style functions. As the question is tagged C++, you may want to use a method instead:
inputMat.copyTo(outputMat, maskMat);
All objects are of type cv::Mat.
Please be aware that the masking is binary. Any non-zero value in the mask is interpreted as 'do copy'. Even if the mask is a greyscale image.
Also be aware that the .copyTo() function does not clear the output before copying.
If you want to permanently alter the original Image, you have to do an additional copy/clone/assignment. The copyTo() function is not defined for overlapping input/output images. So you can't use the same image as both input and output.
You don't apply a binary mask to an image. You (optionally) use a binary mask in a processing function call to tell the function which pixels of the image you want to process. If I'm completely misinterpreting your question, you should add more detail to clarify.
Well, this question appears on top of search results, so I believe we need code example here. Here's the Python code:
import cv2
def apply_mask(frame, mask):
"""Apply binary mask to frame, return in-place masked image."""
return cv2.bitwise_and(frame, frame, mask=mask)
Mask and frame must be the same size, so pixels remain as-is where mask is 1 and are set to zero where mask pixel is 0.
And for C++ it's a little bit different:
cv::Mat inFrame; // Original (non-empty) image
cv::Mat mask; // Original (non-empty) mask
// ...
cv::Mat outFrame; // Result output
inFrame.copyTo(outFrame, mask);
You can use the mask to copy only the region of interest of an original image to a destination one:
cvCopy(origImage,destImage,mask);
where mask should be an 8-bit single channel array.
See more at the OpenCV docs
Here is some code to apply binary mask on a video frame sequence acquired from a webcam.
comment and uncomment the "bitwise_not(Mon_mask,Mon_mask);"line and see the effect.
bests,
Ahmed.
#include "cv.h" // include it to used Main OpenCV functions.
#include "highgui.h" //include it to use GUI functions.
using namespace cv;
using namespace std;
int main(int argc, char** argv)
{
int c;
int radius=100;
CvPoint2D32f center;
//IplImage* color_img;
Mat image, image0,image1;
IplImage *tmp;
CvCapture* cv_cap = cvCaptureFromCAM(0);
while(1) {
tmp = cvQueryFrame(cv_cap); // get frame
// IplImage to Mat
Mat imgMat(tmp);
image =tmp;
center.x = tmp->width/2;
center.y = tmp->height/2;
Mat Mon_mask(image.size(), CV_8UC1, Scalar(0,0,0));
circle(Mon_mask, center, radius, Scalar(255,255,255), -1, 8, 0 ); //-1 means filled
bitwise_not(Mon_mask,Mon_mask);// commenté ou pas = RP ou DMLA
if(tmp != 0)
imshow("Glaucom", image); // show frame
c = cvWaitKey(10); // wait 10 ms or for key stroke
if(c == 27)
break; // if ESC, break and quit
}
/* clean up */
cvReleaseCapture( &cv_cap );
cvDestroyWindow("Glaucom");
}
Use copy with a mask.
Code sample:
Mat img1 = imread(path); // Load your image
Mat mask(img1 .size(),img1 .type()); // Create your mask
mask.setTo(0);
Point center(img1.cols/2, img1.rows / 2);
const int radius = img1.cols / 5; // Circle radio
circle(mask, center, radius, 255, FILLED);// Draw a circle in the image center
Mat img2(img1 .size(),img1 .type()); // Outimage
img2.setTo(0); // Clear data
img1.copyTo(img2, mask); // Only values at mask > 0 will be copied.