Calculating SIFT features

[Ibtissam Bouazzati]

Hi, 

I am beginner in HIPI and the domain of image processing and I have a question about calculating an image features using sift.  For that, I use this code : 

************************** CODE : BEGIN  ****************************

// value is the Float Image

int w = value.getWidth();

int h = value.getHeight();

// conversion of floatImage to opencv Mat

Mat image = new Mat(h,w, CV_32FC3);

FloatIndexer idx = image.createIndexer();

float[] rgb = {0.0f,0.0f,0.0f};

int indice =0;

for (int i =0;i<h;i++) {

    for (int j=0;j<w;j++) {

          rgb[0] = value.getPixelArray().getElem(indice); 

          rgb[1] = value.getPixelArray().getElem(indice+1);

          rgb[2] = value.getPixelArray().getElem(indice+2);

          idx.put(i, j, 0, rgb[0]);

          idx.put(i, j, 1, rgb[1]);

          idx.put(i, j, 2, rgb[2]);

          indice = indice+3;

     }

}

// Convert the image from RGB To grayscale

cvtColor(image, image, COLOR_RGB2GRAY); 

// SIFT features

SIFT sift = new SIFT().create(); 

KeyPointVector keypoints = new KeyPointVector();

Mat descriptors = new Mat();

sift.detect(image, keypoints);

sift.compute(image,keypoints,descriptors);

****************************CODE : END ***************************

Then, I get this error :

Error: java.lang.RuntimeException: /projects/bytedeco/javacpp-presets/opencv/cppbuild/linux-x86_64/opencv_contrib-3.2.0/modules/xfeatures2d/src/sift.cpp:770: error: (-5) image is empty or has incorrect depth (!=CV_8U) in function detectAndCompute

at org.bytedeco.javacpp.opencv_features2d$Feature2D.detect(Native Method)

at org.hipi.examples.Hbasefeatures$FeaturesMapper.map(Hbasefeatures.java:185)

at org.hipi.examples.Hbasefeatures$FeaturesMapper.map(Hbasefeatures.java:63)

at org.apache.hadoop.mapreduce.Mapper.run(Mapper.java:146)

at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:787)

at org.apache.hadoop.mapred.MapTask.run(MapTask.java:341)

at org.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:164)

at java.security.AccessController.doPrivileged(Native Method)

at javax.security.auth.Subject.doAs(Subject.java:422)

at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1698)

at org.apache.hadoop.mapred.YarnChild.main(YarnChild.java:158)

From the error, I think that the problem is that the type of opencv Mat that I give to the detect and compute methods is not the right one.

Can you tell me please if the image conversion from RGB to Grayscale isn't right ? 

Have you any idea about how  this opencv Mat passed to thoses methods should be ? 

[Samuel Audet]

It's just saying that you need an CV_8UC3 image instead of CV_32FC3. We can use Mat.convertTo() for that. 

Samuel

2017/07/12 23:16 "Ibtissam Bouazzati" <bouazzati...@gmail.com>:

--

---
You received this message because you are subscribed to the Google Groups "javacv" group.
To unsubscribe from this group and stop receiving emails from it, send an email to javacv+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Ibtissam Bouazzati]

Exactly, it works ! Thank you for your answer 

Samuel

To unsubscribe from this group and stop receiving emails from it, send an email to javacv+un...@googlegroups.com.