Ask Your Question

ArcadeBit's profile - activity

2015-04-08 09:27:44 -0500 asked a question How does LBP work?

Hello there,

I understand how a Haar-Feature work, because it is easy to understand how you can find and deteect an "Interesting Area" which contains an "unique" Area. The example in the face detection with the Eyebrows is easy to understand for me.

But how does lbp Works? Are there any example how it defines an interesting Area?

Kind Regards

2015-03-17 12:41:45 -0500 received badge  Supporter (source)
2015-03-17 12:41:11 -0500 commented answer FREAK experience requiered

I want to detect the face with the face-detector first and then switch to tracking with FAST, FREAK, SURF, SIFT or BRISK, depending on the performance (speed and accuracy). Can you tell me, how fast your solution was?

2015-03-16 03:38:00 -0500 asked a question FREAK experience requiered

Hello there,

I am working on a facetracking for quite a while now, to control a virtual 3D scene. I run a simple faceDetection on an 320x240 image and track the position with the detected face. Sadly, the face detection isn't rotation invariant, so I createt a new classifier, but the false positive rate was just terrible. So I started to rotate the image (20 degree to the left and right), detected my Face on 3 images parralel in separate Threads, rotate my Keypoint (if requiered) back and done. This kinda works, but the Thread handling isn't nice.

My solutuion is kinda fast. I get my Keypoint in under 10ms, but I don't like the solution. Now I wanted to give FREAK, SURF, SIFT and ORB a chance. How fast are these OpenCV modules? Can anyone tell me a realistic time to detect a face Position in a 320x240 image?

2015-03-16 03:10:34 -0500 answered a question Capture image from Playstation Eye doesn't work.

Yeah, I was just stupid. The driver for the PlaystationEye is a 32 bit driver and doesn't work with the x64 JVM. An switch to the x86 JRE solved the problem.

2015-02-23 15:26:22 -0500 commented question Is there any code for Eye and Yawning detection using opencv?

OpenCV has a HaarCascade to detect Eyes, but it isn't very fast. My Tip: Search for the face with the lbp cascade first, create a sub-image from the upper half of the face and run an eyedetection on that part. If you don't find an eye in 5-10 Images in a row, hit the alarm. 10 images would be closed eyes for 0.33 Seconds.

With a good setup you should detect the Face and eye in less then 25ms, which is still faster then a 30FPS camera.

2015-02-20 18:32:38 -0500 received badge  Editor (source)
2015-02-20 16:49:27 -0500 asked a question Capture image from Playstation Eye doesn't work.

Hello there,

i am working on a Face Detection application in Java with OpenCV. I started with my internal camera and now I want to switch to the Playstation Eye, for the better performance.

I use

VideoCapture imageGrabber = new VideoCapture(0);;

to read an image.

With my internal Webcam everything works fine. But: OpenCV doens't open the Playstation Eye, even when i deactivate my interal camera.

I use the driver from Code Laboratories. I looked into the code from JavaCV. JavaCV use a library from Code Laboratories to access the camera. Doesn't the camera work with OpenCV?

2015-02-20 16:23:00 -0500 commented question Wrong arguments for CascadeClassifier detection

You are missing a Size at the end. Should look like this:

face_cascade.detectMultiScale(mat, faces, 1.1, 2, 0, Size(30, 30), Size(500,500));

2015-02-13 02:52:14 -0500 received badge  Enthusiast
2015-02-08 12:41:32 -0500 asked a question OpenCV and all about speed

I am currently working on a face detection application.

I use the integrated webcam from my laptop. The maximum framerate of the webcam is 30FPS but I only get 15FPS out of the webcam. Measuring it, I take 30 Pictures in 2 Second or 55-75ms per picture.

Would a switch to a faster webcam, like the Playstation Eye, improve the performance?

The other questions are about the Face Detector itself.

    Size minSize = new Size(50, 50);
    Size maxSize = new Size(500, 500);
    faceDetector.detectMultiScale(image, result, 1.05, 3, CASCADE_FIND_BIGGEST_OBJECT, minSize, MaxSize);

The first is about the scaleFactor. What is the maximum value, which still achieves good results. I want to have a performance orientated version.

The second one is about the Flag. I use CASCADE_FIND_BIGGEST_OBJECT. Does this change the algorithm to start with a big sub-window and start to scale down, until it finds a Object and stop? Or does it still run through the complete pyramid from sub-windows?

My last question is about the multicore use from OpenCV. Does it use more then one core to detect a face or does it only run on one core. If it only runs on one core, is it smart to start 2 - 3 Threads, each one running a face detection with different Options? I would use a fast one, but with a bad detection and a slower one with a better detection. Would this make sense?