1. Forum
  2. Usenet
  3. SCI.IMAGE.PROCESSING

  • Image processing factors

    From Ken Wei@21:1/5 to All on Sun May 29 03:08:28 2016
    Base on recently study, most of the image processing is base on pixel by pixel, it's hard for machine to know where is the cat. because for a machine could only process 0 or 1. But the study our human's way to detect the object or the key component
    inside your view, is base on experience.

    for example, when you are inside a lift, you can fast to know the buttons are either on left or right, and the LED screen on on top. When you arrive at home, you will know door switch is around the middle side of the door.

    This is a learning processing, which means, the experience and the scene of the life should be one of important for us to detect specific object in normal life.

    But when a new image was gave to us, we still need to detect the image from left to right, up to down to learn it.

    Experience need to be considered inside image processing, at least you know where you are, you could easy to find what you want.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Dale@21:1/5 to qianjun.w@gmail.com on Sun May 29 14:59:25 2016
    On Sun, 29 May 2016 03:08:28 -0700 (PDT), Ken Wei
    <qianjun.w@gmail.com> wrote:

    Base on recently study, most of the image processing is base on pixel by pixel, it's hard for machine to know where is the cat. because for a machine could only process 0 or 1. But the study our human's way to detect the object or the key component
    inside your view, is base on experience.

    for example, when you are inside a lift, you can fast to know the buttons are either on left or right, and the LED screen on on top. When you arrive at home, you will know door switch is around the middle side of the door.

    This is a learning processing, which means, the experience and the scene of the life should be one of important for us to detect specific object in normal life.

    But when a new image was gave to us, we still need to detect the image from left to right, up to down to learn it.

    Experience need to be considered inside image processing, at least you know where you are, you could easy to find what you want.

    there is also appearance matching versus color matching like CIECAM02
    --
    Dale
    http://www.dalekelly.org

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)