jms's profile - activity

%04d.tiff looks for 0000.tiff, 0001.tiff, 0002.tiff, 0003.tiff and 0004.tiff. It does not find for example 0100.tiff

Sure I would like to write a module for combining multiple xml models, but in python a LBPHFaceRecognizer cannot read fr

I would like to write a module for combining multiple xml models, but in python a LBPHFaceRecognizer cannot read from a

Merging / Combining separately trained LBPH models If I have trained a LBPHFaceRecognizer for different faces and stored

Is there a way to get the confidence value without building the code? In the ubuntu repos there is only openCV 3.1.0

If I read a mp4 video with SAR 4:3 and DAR 16:9 and I store or replay video frames without knowing the DAR, the DAR is lost and the video frames are displayed at 4:3 and so are distorted.

playing the video with VLC it appears as 16:9 and looks good.

In python I process a mp4 video with

stream = cv2.VideoCapture(filename)

The video has a resolution of 1440x1080 and a SAR of 4:3, but a display aspect ratio (DAR) of 16:9. This is how it comes from my sony camera. To preserve the display aspect ratio, I need to know it. How can I get the display aspect ratio with OpenCV? ffmpeg -i tells me this:

Stream #0:0(und): Video: h264 (Main) (avc1 / 0x31637661), yuv420p(tv, bt709/bt709/iec61966-2-4), 1440x1080 [SAR 4:3 DAR 16:9]

I'm looking for a pre trained haar cascade file for detecting digits (not hand written). I thought there must me hundreds in the net but I found only https://github.com/joyeecheung/digit-... which does not work well.