Training method and device of video processing model and video processing method and device

Abstract

The embodiment of the invention provides a video processing model training method and device and a video processing method and device, and the method comprises the steps: inputting each frame of image of sample video data into a preset single-frame model, and obtaining a first recognition result of a target object in each frame of image outputted by each network layer; according to the time sequence of the images in the sample video data, inputting each frame of image into a to-be-trained time sequence model to obtain a second recognition result of the target object in each frame of image output by each network layer; comparing the first recognition result and the second recognition result of each frame of image output by each network layer, and determining the total loss value of the time sequence model recognition image; if it is determined that the time sequence model converges based on the total loss value, ending training. By applying the technical scheme provided by the embodiment of the invention, the problem that the cost of service resources is greatly increased due to the fact that a large amount of approximately repeated calculation is carried out on video processing because inter-frame redundancy exists between the images of the video data is solved.

Description

technical field

[0001] The present disclosure relates to the technical field of video processing, in particular to a video processing model training method and device, and a video processing method and device. Background technique

[0002] In the video processing scenario, many intelligent applications need to deploy a deep neural network model on the server or mobile terminal, and use the deep neural network model to output a result for each frame of video data. For example, when using magic expression special effects, it is necessary to use the deep neural network model to obtain the key points of the face on each frame of video data; when using the background blur special effect, it is necessary to use the deep neural network model to obtain each The segmentation mask (mask) of the foreground and background of the person on the frame image; when performing adaptive video coding and compression, it is necessary to use a deep neural network model to obtain the mask of the s...

Examples

Embodiment Construction

[0068] In order to enable ordinary persons in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings.

[0069] It should be noted that the terms "first" and "second" in the specification and claims of the present disclosure and the above drawings are used to distinguish similar objects, but not necessarily used to describe a specific sequence or sequence. It is to be understood that the data so used are interchangeable under appropriate circumstances such that the embodiments of the disclosure described herein can be practiced in sequences other than those illustrated or described herein. The implementations described in the following exemplary examples do not represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatuses and methods consi...

technical field

[0001] The present disclosure relates to the technical field of video processing, in particular to a video processing model training method and device, and a video processing method and device. Background technique

[0002] In the video processing scenario, many intelligent applications need to deploy a deep neural network model on the server or mobile terminal, and use the deep neural network model to output a result for each frame of video data. For example, when using magic expression special effects, it is necessary to use the deep neural network model to obtain the key points of the face on each frame of video data; when using the background blur special effect, it is necessary to use the deep neural network model to obtain each The segmentation mask (mask) of the foreground and background of the person on the frame image; when performing adaptive video coding and compression, it is necessary to use a deep neural network model to obtain the mask of the s...

Examples

Embodiment Construction

[0068] In order to enable ordinary persons in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings.

[0069] It should be noted that the terms "first" and "second" in the specification and claims of the present disclosure and the above drawings are used to distinguish similar objects, but not necessarily used to describe a specific sequence or sequence. It is to be understood that the data so used are interchangeable under appropriate circumstances such that the embodiments of the disclosure described herein can be practiced in sequences other than those illustrated or described herein. The implementations described in the following exemplary examples do not represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatuses and methods consi...

Images