| CPC G06V 20/44 (2022.01) [G06T 5/002 (2013.01); G06T 5/40 (2013.01); G06T 7/10 (2017.01); G06V 10/82 (2022.01); G06V 10/945 (2022.01); G06V 20/46 (2022.01); G06V 20/49 (2022.01); G06V 20/52 (2022.01); G06T 2207/10016 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30232 (2013.01)] | 20 Claims |
|
1. A system for detecting one or more anomalous events in video data, the system comprising:
a processor; and
a machine-readable medium in operable communication with the processor and having instructions stored thereon that, when executed by the processor, perform the following steps:
a) receiving the video data;
b) performing a noise cleansing on the video data to provide cleansed video data;
c) performing feature extraction on the cleansed video data using a neural network to give feature-extracted video data;
d) performing instance segmentation on the feature-extracted video data to give a plurality of segmented instances, each segmented instance of the plurality of segmented instances representing a predetermined amount of time of the feature-extracted video data;
e) performing instance summation on the plurality of segmented instances by calculating a sum of extracted feature values on each segmented instance of the plurality of segmented instances, to give a plurality of segmented instance sums;
f) performing instance difference calculation on the plurality of segmented instance sums to determine a plurality of difference values for the plurality of segmented instance sums, respectively;
g) normalizing the plurality of difference values to give a plurality of normalized difference values, each normalized difference value of the plurality of normalized difference values representing how different each instance is to at least one adjacent instance; and
h) if the normalized difference value for a given instance is greater than a predetermined threshold, marking the given instance as an anomalous event.
|