Highlight detection method and device based on image segmentation, terminal and storage medium

Abstract

The invention discloses a highlight detection method and device based on image segmentation, a terminal and a storage medium, and the method comprises the steps: obtaining a plurality of image frames in a to-be-detected image, and determining highlight feature vectors corresponding to the plurality of image frames; segmenting the plurality of image frames according to the highlight feature vectors respectively corresponding to the plurality of image frames to obtain a plurality of image segments and average highlight feature vectors respectively corresponding to the plurality of image segments; and determining a to-be-detected image frame in the to-be-detected image, determining a target image segment corresponding to the to-be-detected image frame from the plurality of image segments, and performing highlight detection on the to-be-detected image frame according to the average highlight feature vector corresponding to the target image segment. The problem that a highlight detection method in the prior art usually takes a single static image and a specific background in the image as processing objects, and complex and dynamic scenes related to sequence images or videos shot in a moving mode are difficult to process is solved. 本发明公开了基于影像分段的高光检测方法、装置、终端及存储介质，通过获取待检测影像中的若干影像帧，确定若干影像帧分别对应的高光特征向量；根据若干影像帧分别对应的高光特征向量对若干影像帧进行分段，得到若干影像段和若干影像段分别对应的平均高光特征向量；确定待检测影像中的待检测影像帧，从若干影像段中确定待检测影像帧对应的目标影像段，根据目标影像段对应的平均高光特征向量对待检测影像帧进行高光检测。解决了现有技术中的高光检测方法通常是以单张静态影像与影像中的特定背景作为处理对象，难以处理移动拍摄的序列影像或视频涉及的复杂、动态的场景的问题。