Ask Your Question

Dekay's profile - activity

2017-10-02 11:42:40 -0600 received badge  Student (source)
2017-02-23 06:05:24 -0600 asked a question Where can I find the minimum supported Mac OSX versions supported by each version of opencv?

Hi Folks,

tl;dr: App works fine on my osx system, but crashes on older ones. I am looking for information on the minimum required osx version for opencv library versions, expecially 2.4.13 and 3.1.

The prelude: I am writing a face recognition application for mac osx. I develop with xcode 8.2 on mac osx 10.11.6 and use opencv 3.1 libraries. I included the necessary opencv libraries in my app bundle and relinked the running search paths appropriately so the linker can find the libraries at the correct locations. On my machine everything works fine.

The problem: Now I tried using my app on mac osx 10.8. The app does not start and gives me the following error:

Dyld Error Message: Symbol not found: ___sincos_stret Referenced from: /Users/USER/Downloads/myApp.app/Contents/MacOS/../Frameworks/opencv/libopencv_imgproc.3.1.dylib Expected in: /usr/lib/libSystem.B.dylib

The library in question (libSystem.B.dylib) is a system library which exists on that machine. After some research I found out that the symbol sincos_stret is a funcion that is only exposed after mac osx 10.9, so that is probably why it does not find it on 10.8.

The arising question: After my problem I searched for information on the necessary osx versions both on the opencv website and on the internet in various forums but did not find any specifications. This astonishes me, since it is a very valuable information to have. Does anybody know where to get it. I would like to avoid fixing this problem after alot of work, only to find out that another problem appears because of compatibility problems of opencv and the osx versions.

Thanks for your answers and kind regards, David

2016-11-23 08:30:51 -0600 asked a question Does openCV provice anti spoofing techniques for face detection?

Hi guys,

I am currently working on a face recognition project with openCV. I managed to set up the face detection and recognition already. Now I am looking for solutions to discriminate between "real faces" and "spoofed faces". With spoofed faces I refer to photos / vidoes / masks of a face, that are held in front of my camera to trick the recognition system to be someone else.

I browsed the openCV reference page for some time now, but sadly it does not seem to provide such functionality. If anyone of you knows some off-the-shelve solutions, please share them with me.

Thanks in advance!