pavrom's profile - activity

This question was answered here http://stackoverflow.com/questions/15024259/noisy-kinect-color-rgb-images. Thank you.

i'm using opencv interface (http://docs.opencv.org/doc/user_guide/ug_highgui.html?highlight=kinect) to get color (rgb) and depth frames from a kinect camera. For a standard VGA 640x480 resolution and with code like

capture.retrieve( bgrImage, OPENNI_BGR_IMAGE );

i get this :

I think this is really noisy. Is this normal quality for a kinect rgb camera? I tried various filtering (blurring, sharpening, opening..) procedures but i got minor improvements.

I'd like to do something like this to get bow descriptors :

    Ptr<FeatureDetector > detector(new SurfFeatureDetector());
    Ptr<DescriptorMatcher > matcher(new BruteForceMatcher<L2<float> >());
    Ptr<DescriptorExtractor > extractor(new OpponentColorDescriptorExtractor(Ptr<DescriptorExtractor>(new SurfDescriptorExtractor())));
    Ptr<BOWImgDescriptorExtractor> bowide(new BOWImgDescriptorExtractor(extractor,matcher));
    bowide->setVocabulary(vocabulary);

However, i'm extracting different kinds of descriptors for rgb/depth couples, (SURF for color images and VFH for depth ones), so my question is : how can i define a custom class which takes two images and computes the descriptors (vectors of 128+308 elements) to be used with the above code?

Thank you.

@FrieS: thank you very much. Right now i'm trying to implement the strategy outlined in one of my previous comments (Feb 5 '13 - right below). I'd like to know your opinion on this. Also, i didn't know about HOD, thanks for that too..

Thank you. For the vector-size problem i could use PCA and take only n components. The dataset i'm using was collected using OpenNI, which has a flag for registering the RGB and depth images so that they are aligned. Also, right now i'm more interested in recognition than detection. I'll probably try both strategy (one and two classifiers).

Thank you. I was thinking about collecting multiple features following the bag of words approach. For visual appearance any fast-computation feature (for example, FREAK) could be useful while for depth i was thinking about using fast point feature histograms (FPFH) (http://pointclouds.org/documentation/tutorials/fpfh_estimation.php#fpfh-estimation). After collecting both kind of features for each training sample, i would like to get the descriptors to train a single classifier. Do you think this strategy could be feasible? (I'm actually kind of a newbie to the all thing).

For a task of object recognition/detection, i'm looking for a way to employ depth maps i got from a kinect style camera to increase accuracy of an eventual classfier. I was wondering if computing standard features (for example, SIFT, SURF, FAST or HOG) on depth maps could be useful in this context.

Thank you.