I meant a local change in luminosity.
The output is just one bit. These methods detect changes, movements and combinations of both. One bit does not give enough information to be able to distinguish.
I meant a local change in luminosity.
The output is just one bit. These methods detect changes, movements and combinations of both. One bit does not give enough information to be able to distinguish.
Yes. if some spot brightens, that will be considered a change. However, given uniform lighting, that doesn't sound like a big problem.
I would say b. CRC typically outputs multibit differences for 1 bit disturbances in the input field.
In article , joseph2k wrote: [... CRC for video motion detection ...]
The trick is to make sure the CRC gets the same string of bits each time unless something moved. The trouble is coming up with a way to ensure that.
The same threshold Luhan proposed to use after his sums can be used before the CRC. So, yes, "b."
Best part is we need neither sums nor CRCs. See my other post.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required