5 Replies Latest reply on May 1, 2011 5:37 PM by Drew747

    Face recognition

    Francisc Level 3

      Hello,

       

      Can you point me into the right direction?

      I am trying to build a face recognition application that can track a person's face over video footage by scanning it frame-by-frame.

      I want it to be an AIR app built with Flex 4.5.

       

      Basically, I need a hint as to how scanning an image for another image works in AS3.

       

      Thank you.

        • 1. Re: Face recognition
          UbuntuPenguin Level 4

          I'm not sure this is the right forum to be posting this question in.  Your problem is composed of MANY sub-problems.  The first few being, what are you video acquisition methods.  How are you "slicing" the video into frames.  And then there is the part that people with PhDs are paid to do, what is your algorithm for finding and tracking the face going to be.

          • 2. Re: Face recognition
            Francisc Level 3

            Hey,

             

            It is a complex problem, but the part I am `really` curious about is how to match an image inside another image in AS3.

            • 3. Re: Face recognition
              Level 2

              don't use AS3. acquire the image and send it to a server for heavy lifting with an appropriate language and routines already available in C# for instance.

               

              it's like the guy that wanted to write image editing software and kept running into memory issues with flash player. don't put yourself thru it in AS3.

              • 4. Re: Face recognition
                Francisc Level 3

                It is supposed to be a desktop application, so the server isn't a problem.

                If it isn't `doable` in AS3, I will give other languages a thought.

                • 5. Re: Face recognition
                  Level 2

                  Why as3 cuz of the camera interface  What I was suggesting is that u capture then send it to a server that has powerful progr languages  It is like voice recogn on Android. The sound captured is sent to the cloud for interpretation and the answer is sent back  I don't know what your constraints and flexibilities are so I have little to go on