See On the Sensor Pattern Noise Estimation in Image Forensics: A Systematic Empirical Evaluation.pdf table III.
Using many images seem to improve the PRNU estimation but in justice cases we do not have tens of images, do we? Could think that images not criminal related but regular images could do the job but then most of the estimated PRNU is not for the criminal related case, so does not seem correct.
See `On the Sensor Pattern Noise Estimation in Image Forensics: A Systematic Empirical Evaluation.pdf` table III.
Using many images seem to improve the PRNU estimation but in justice cases we do not have tens of images, do we? Could think that images not criminal related but regular images could do the job but then most of the estimated PRNU is not for the criminal related case, so does not seem correct.
Correlation just gives between absolutely 0 and 1 if both signal match but may have the same correlation for a single image as for a whole dataset so can take into account the quantity to define the confidence?
Correlation just gives between absolutely 0 and 1 if both signal match but may have the same correlation for a single image as for a whole dataset so can take into account the quantity to define the confidence?
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
See
On the Sensor Pattern Noise Estimation in Image Forensics: A Systematic Empirical Evaluation.pdftable III.Using many images seem to improve the PRNU estimation but in justice cases we do not have tens of images, do we? Could think that images not criminal related but regular images could do the job but then most of the estimated PRNU is not for the criminal related case, so does not seem correct.
Correlation just gives between absolutely 0 and 1 if both signal match but may have the same correlation for a single image as for a whole dataset so can take into account the quantity to define the confidence?