Method and device for detecting specific object in video data and computing device

Abstract

The invention discloses a method and device for detecting a specific object in video data and a computing device. The method includes the following steps: the video data is obtained; the video data isinput into a second neural network to obtain a detection result of the video data; the second neural network is obtained after output data of at least one intermediate layer of the pre-trained firstneural network is subjected to guiding and training operation, and the first neural network is greater than the second neural network in terms of layer quantity; label attribute information of the video data is determined according to the detection result of the video data. The video data is marked according to the label attribute information of the video data. According to the method disclosed inthe invention, the neural network with a small number of layers is used for rapidly and accurately detecting the video data, a detection result is obtained, the label attribute information of the video data can be conveniently determined, and therefore the video data can be accurately marked.

Description

technical field [0001] The invention relates to the field of deep learning, in particular to a method and device for detecting a specific object in video data, and a computing device. Background technique [0002] With the development of network technology, users can obtain massive video data from the Internet. For users, they only hope to obtain the video data they are interested in from the massive video data. Take fans of a star as an example, fans only pay attention to the video data of the star; for companies or families that install security monitoring, they pay attention to whether there are other strangers in the monitored video data. [0003] In the prior art, video data can be checked manually to determine whether the video data contains a specific object. However, the manual review method requires a large amount of manpower, which increases the review cost, and the review efficiency is low. Or in the prior art, a neural network is used to detect video data. Ho...

AI Technical Summary

Problems solved by technology

However, the generally used neural network often has multiple layers of intermediate layers, which can obtain accurate detection results, but the calculation speed of multiple layers of intermediate layers will be slow, and the video cannot be detected quickly, and the detection results cannot be fed back to the user in time.

Moreover, users use more mobile devices when acquiring video data. The neural network that mobile devices can run is generally a neural network with fewer intermediate layers. Since the number of intermediate layers is small, its calculation speed is faster, but it is affected by The limitation of the number of layers may cause problems such as limited computing power, poor fitting ability, and inaccurate results.

Images