Oct-16-2022, 05:27 PM
Hello,
many thanks to all the hard-working people, who shared their knowledge!!
Reading this, I've got a starting point to work with.
I will notify you if I need assistance...
Thanks a lot...
-------------------------------------------------------------------
Just for explanation (because the question arised what I'm planning to do with the list of hash codes):
Digitizing Super8-Movies for customers is a part of my work. During the recording process the film can hang. These hanging sequences I have to cut out in a video editing software.
It is an effort to watch the digitzed film and check for hangers.
What I'm planning to do is the following:
First, I extract the frames from the digitized film.
Then, I want to examine each frame with perceptual hashing.
To every frame it belongs a hash code.
When the film hangs, the same frame is digitized again and again...
Because of this it comes to an equal sequence of the same hash code.
Knowing the start and the end of a hanger I will calculate the position in film (hh:mm:ss) of the hanger.
I think that perceptual hashing is more suitable to examine the frames, because it isn't too accident-sensitive against small changes in the frame.
I tested an other method and changed only one pixel in the same frame and the changed frame was identified as a different picture.
Because of this I'm thinking that the perceptual hashing is better suitable for detecting the same frame (->hanger).
many thanks to all the hard-working people, who shared their knowledge!!
Reading this, I've got a starting point to work with.
I will notify you if I need assistance...
Thanks a lot...
-------------------------------------------------------------------
Just for explanation (because the question arised what I'm planning to do with the list of hash codes):
Digitizing Super8-Movies for customers is a part of my work. During the recording process the film can hang. These hanging sequences I have to cut out in a video editing software.
It is an effort to watch the digitzed film and check for hangers.
What I'm planning to do is the following:
First, I extract the frames from the digitized film.
Then, I want to examine each frame with perceptual hashing.
To every frame it belongs a hash code.
When the film hangs, the same frame is digitized again and again...
Because of this it comes to an equal sequence of the same hash code.
Knowing the start and the end of a hanger I will calculate the position in film (hh:mm:ss) of the hanger.
I think that perceptual hashing is more suitable to examine the frames, because it isn't too accident-sensitive against small changes in the frame.
I tested an other method and changed only one pixel in the same frame and the changed frame was identified as a different picture.
Because of this I'm thinking that the perceptual hashing is better suitable for detecting the same frame (->hanger).