| CPC G11B 27/34 (2013.01) [G06F 18/22 (2023.01)] | 18 Claims |
|
1. A method of processing video, the method comprising:
receiving content comprising a plurality of segments; and
determining a terminus time of a segment of the plurality of segments based on an audio spectrogram associated with the content, the audio spectrogram comprising pixels with brightness values over a plurality of frequencies along a first axis and over a plurality of timepoints along a second axis,
wherein determining the terminus time of the segment comprises:
determining a first estimate of the terminus time by:
resizing the audio spectrogram such that, for each timepoint of the plurality of timepoints, a single brightness value is generated to represent corresponding brightness values at each timepoint;
for a subset of timepoints of the resized audio spectrogram:
positioning a filter with respect to the resized audio spectrogram such that a head of the filter aligns with a first timepoint of the subset of timepoints, wherein the filter comprises values and is associated with a window size; and
calculating a measure of similarity between the values of the filter and the single brightness values corresponding to the first timepoint through a second timepoint as determined based on the window size; and
determining the first estimate as corresponding to a timepoint of the subset at which a greatest measure of similarity is calculated.
|