2017-06-15 01:53:30 -0600 | asked a question | LUCID does not work I am extracting keypoints using StarDetector. I am trying to compute descriptors using LUCID but I am getting the following error. Can anyone please explain why I am getting this error?
|
2017-06-12 23:34:50 -0600 | commented answer | Modifying inpainting to achieve proper blur Create a mask only for facial features such as eyes. Not the entire face. Creating for the entire face does not guarantee proper blurred regions for inside the face. |
2017-06-12 05:15:03 -0600 | answered a question | Modifying inpainting to achieve proper blur This cartoonised image does not work as a proper mask. You need to create a mask for your eyes and other facial features if you want to remove them. Create a binary mask (maybe you can use MS Paint or for more accurate results you can use Photoshop) painting your facial features black and the rest as white. Pass it as the second argument and it will blur out all your facial features. Hope it helps :)! This is the photo from which I want to blur the kid. This is the mask I use to blur it. This is the final result I get on using a binary mask. |
2017-06-12 01:38:20 -0600 | asked a question | Unable to run programs of SFM (Structure From Motion) module I am getting an error when I am trying to execute a sample code in SFM module. Initially SFM was not present in my contrib directory. So i downloaded the latest contrib and pasted the sfm folder in my contrib directory. I used the CMakeLists.txt that was present in the sfm module. Then I tried to build it in my own directory with scene_reconstruction.cpp and the CMakeLists.txt which I took from the sfm module. Following is the error which I am getting PS: I went through all the required downloads of the libraries mentioned in the SFM documentation page before starting off. |
2017-06-07 23:20:08 -0600 | commented question | The hausdorff distance is not zero even when 2 same images are passed Yes. I am using jpg images. If not jpg, what images should I use? |
2017-06-07 23:18:48 -0600 | commented answer | The hausdorff distance is not zero even when 2 same images are passed No. I have done it in openCV. The value is less, say a single digit, but is never zero. |
2017-06-07 05:11:51 -0600 | received badge | ● Scholar (source) |
2017-06-07 05:11:45 -0600 | answered a question | The hausdorff distance is not zero even when 2 same images are passed The hausdorff considers the worst matchin pairs in the two images. Hence it never returns a zero value. Whereas the shapeContextDistanceExtractor returns a zero value when 2 same images are passed. |
2017-06-07 05:09:38 -0600 | received badge | ● Supporter (source) |
2017-06-07 02:29:27 -0600 | answered a question | Can't seem to use OpenCV Hello. As per your question, I guess that you have installed OpenCV but are having problem while compiling and running the code. Here are steps you should follow. 1. Name your .cpp file as main.cpp 2. Create a .txt file and name it as CMakeLists.txt 3. In the CmakeLists.txt file paste the following lines " cmake_minimum_required(VERSION 2.6) cmake_policy(SET CMP0012 NEW) PROJECT(project) #Requires OpenCV FIND_PACKAGE(OpenCV 3.0.0 REQUIRED) MESSAGE("OpenCV version : ${OpenCV_VERSION}") include_directories(${OpenCV_INCLUDE_DIRS}) link_directories(${OpenCV_LIB_DIR}) ADD_EXECUTABLE(app main.cpp) TARGET_LINK_LIBRARIES(app ${OpenCV_LIBS}) " without the doublequotes at the start and the end. Save the file.
|
2017-06-06 05:39:52 -0600 | answered a question | Remove noise image? You can use fastNlMeansDenoising or fastNlMeansDenoisingColored function in OpenCV to remove gaussian white noise or colored noise respectively. Hope it helps. |
2017-06-06 01:30:59 -0600 | received badge | ● Necromancer (source) |
2017-06-06 01:22:01 -0600 | answered a question | denoise_TVL1 not working? Hello. It seems you are passing one image in I. But the first argument for denoise_TVL1 is a vector of images. Try passing vector of images rather than a single image. Hope it helps. |
2017-06-02 00:57:27 -0600 | asked a question | The hausdorff distance is not zero even when 2 same images are passed I am passing 2 same images in hausdorff and shapeContextDistanceExtractor to calculate distance. We are expecting the answer to be zero since the images are same but we are getting a non-zero value. Can somebody please explain why.
|
2017-06-02 00:47:30 -0600 | received badge | ● Enthusiast |
2017-05-30 01:39:46 -0600 | received badge | ● Editor (source) |
2017-05-29 23:49:50 -0600 | asked a question | In the Histogram Cost Extractor module, what does the values returned by buildcostMatrix function signify? I am using buildcostMatrix function and passing it 2 arguments which are the descriptors of two images. I am getting a matrix of double. But it is very huge and therefore hard to comprehend. Can anyone please explain how to make sensible information out of the huge matrix which is returned by the function? void buildCostMatrix (InputArray descriptors1, InputArray descriptors2, OutputArray costMatrix) |
2017-05-24 23:44:45 -0600 | asked a question | How can I use the NormHistogramCostExtractor class? What function do I need to call to extract the cost? What are the arguments to be passed ? What does norm based cost extraction mean? Do I have to normalize the histogram and then pass the normalized histograms for cost extraction as in EMDL1 or anything else? Please provide suggestions. |
2017-05-24 23:32:27 -0600 | commented answer | How is the EMDL1() function different from EMD() function in OpenCV? Thanks for the explanation. It was really helpful. |
2017-05-24 06:30:26 -0600 | asked a question | How is the EMDL1() function different from EMD() function in OpenCV? What is the difference between EMDL1() and EMD() function in OpenCV? What does 'L1' signify? Please provide suggestions. |