The invention relates to an
online teaching highlight extraction method, and the method comprises the steps: constructing a voice feature
library, setting a highlight criterion, and making a voice feature
library according to the criterion; extracting a highlight part, analyzing a lesson media film, and obtaining an audio original file; carrying out the preprocessing of the audio file, eliminating
silence information, and obtaining a plurality of audio segments; carrying out the
feature matching of each
audio segment with the information in a voice segment
library corresponding to a lesson: carrying out the marking recording if the matching is successful; extracting an
audio segment at a certain length before each marking
record according to the marking recording, forming highlight segments, splicing the highlight segments, and forming a final media file. The method provided by the invention achieves the refining of the feature library according to the industry scene of an application, and can effectively improve the accuracy of
feature matching. Meanwhile, the method is used for the
feature matching of specific objects, so the method also greatly improves the matching efficiency.