..:: ZOMBiE Destinee ::..

Friday, September 30, 2005
Raven Vision Algorithm
[ZOMBiE CyGiG] ranted at 6:21 PM --- Post#112807569883309252

Ok now here is the algorithm I used for searching of people. Normally I like to walk about in the canteen scanning carefully for any kind of people that I do not wish to meet or to look for my class. The Raven Vision is a fast method to search for people in crowded place.

Firstly, keep thinking in your mind the person that you wish to search for. Then we will glance at the scene from left to right. Not that only your head move while your eyes are still. First we mentally separate the people from the background. Meaning your focus will jump from person to person, ignoring the distracting background.

Next as we look through those people very quickly, we must look out for obvious filters. Like if your friend has a spiky hair style, a scan around the canteen can easily ignore those people with other kind of hair style. While if your target is a guy, girls can be ignored.

Then we will look for special features like height or skin tone or special hair bands. Since we are keeping an “image” or “profile” of them in our mind, our brain will alert us if we had seen something similar, then you will slow and check carefully if he/she is your target. All these can be done very fast as our brain can alert us very quickly if we spot something similar to what we want.

Then if the target is still not found, we can run a search from right to left again. This time we look for facial features and compare them with his/her image in our brain. This may take a slightly longer time but the search is more detailed.

Run the search in opposite direction once more if your want in greater detail.

Also, we can move about as we scan around, to increase our scope.

This method can be altered to suit our needs, we can scan the room for familiar faces and remember them to make known of who is present in the room. In this case, we do not stop scanning when we had found our target, but rather remember him./her and update on his/her position if your search happen to encounter him/her again.

Here is the algorithm:

LOAD target IMAGE into BRAIN ACTIVE MEMORY (keep thinking of him)
Found = = false
SCAN from LEFT to RIGHT
{
     FILTER search.
     LOOK for obvious characteristic (e.g. hairstyle, height, bag, walking pattern)
     IF ACTIVE MEMORY triggered
{
          COMPARE FACE features with IMAGE
          IF (FACE = = IMAGE)
               Found = true
     }
}

IF (Found! = true)
{
     SCAN from RIGHT to LEFT
     {
EXTRACT FACE from whole body
COMPARE FACE with IMAGE
IF (FACE = = IMAGE)
               Found = true
}
}

View/Post Comments (0)