|
REPORT COPYRIGHT VIOLATION IN REPLY
|
|
Message Subject
|
Someone told me the challenger astronauts are still alive and using their real names.
|
|
Poster Handle
|
Anonymous Coward |
|
Post Content
|
...
Ofcourse you would respond with some bullshit.
YOU ARE A SHILL
Look them up and find the pictures for yourself....
They ARE the same people, there is NO DENYING IT.
Quoting: Anonymous Coward 73650830 Wrong. You're really not going to like the video I'm working on. I did look them up, then I ran their photos through facial recognition. It's able to recognize people like Joe Biden based purely on images taken about 40 years ago or more, and it can recognize younger photos of the Challenger crew, but the people identified as present-day Challenger crew are not the same people. This is an objective, verifiable way to deny it. You can shove your biased and blind tech up your ass. That's just it, it's totally unbiased by its very nature, it's blind to your bias as well as mine. And no, I think I'll shove it up your ass instead. Unless you are using the latest methodologies of facial recog, then it can still be flawed. Are you using ConvNet? If not, then you must be using generalized matching face detection method or adaptive regional blend method. Both are flawed and will not give you a 100% match. They can give you a match within a range that is better than 70% or scaled .70, but it can still be argued depending on the target data how accurate the match is. How many features are you tracking for the face? Front or side profile? How large is your source training data? What is considered a match in terms of how many features being tracked? Are you running multiple inputs and comparing results of all output? Are you running multiple inputs and combining results? How are you tracking outliers? Are you using landmarks to identify facial features? How are you identifying a face after processing? Does each landmark have a variance in accuracy? If so, what is the variance? I could go on and on and on and on. There are many variables to facial recog. You are offering nothing but your word on this and presenting a blackbox saying that it is conclusive. Unless you are willing to have someone provide a code review of your model and verify your claims by running a sample set of data then you cannot claim what you are stating as conclusive. You can, however, claim a high probability. I'm not attacking you. But I am pointing out that we as the ACs cannot take what anyone says at face value.
|
|
|
Please verify you're human:
|
|
|
Reason for copyright violation:
|
