| CPC H04N 21/4542 (2013.01) | 19 Claims |
|
1. A computer-implemented method comprising:
analyzing, using at least one processor, first redacted data that is image or video data, and the first redacted data including first redactions sized and shaped to conceal at least some features of at least first and second objects that:
appear concurrently in at least one stored image corresponding to the first redacted data, or
appear concurrently during at least part of stored video corresponding to the first redacted data, and
wherein:
the analyzing includes coming to a determination that, notwithstanding the at least some features are being concealed, a visible association between the first and second objects in combination with first size and shape definitions for the first redactions is such that an identity deducibility condition is either met or is not met, and
the identity deducibility condition being met or not met following from carrying out of an automatic assessment in which at least one learning machine is employed, and an outcome of the automatic assessment being impacted at least in part by at least one of a relative height percentage difference and a relative width percentage difference between at least two of the first redactions;
when, following the automatic assessment, the identity deducibility condition is not met, generating an indication, using the at least one processor, that the first redacted data is sufficiently redacted for an intended purpose; and
when, following the automatic assessment, the identity deducibility condition is met, converting the first redacted data into second redacted data using the at least one processor, and wherein:
the second redacted data includes second redactions having second size and shape definitions different than the first size and shape definitions, and
the second redactions, as compared to the first redactions, more sufficiently conceal identities of the first and second objects.
|