A Real-time Continuous Frame Information Implantation Recognition System

Abstract

The invention discloses a real-time continuous frame information implantation recognition system, comprising: a video processing module, the video processing module including a frame extraction unit, a feature labeling unit, an implantation information generation unit and a file generation unit; and recognition and presentation module, the identification and presentation module acquires the descriptive file package, the identification and presentation module includes a video acquisition unit, an image identification unit, an implant information acquisition unit and a presentation unit, wherein the video acquisition unit acquires video images in real time information and display it on the display screen, the image recognition unit matches the video image acquired by the video acquisition unit based on the identification information of the key frame in the descriptive file package, identifies the key frame, and the implanted The information acquisition unit acquires annotation information and implantation information corresponding to the key frame, the presentation unit adds annotations to the video image acquired by the video acquisition unit based on the annotation information, and then presents the implantation information on the display screen .

Description

technical field [0001] The invention relates to the field of image processing, in particular, the invention relates to a real-time continuous frame information implantation recognition system. Background technique [0002] At present, the main continuous frame information implantation and recognition methods are mainly implanted in the way of front post, insert post, corner mark association, and pause pop-up, and the user clicks to enter the expressed information page to present. This method can be implanted The amount of information is small, and it takes up video playback time and playback interface, which brings poor user experience. [0003] Other existing continuous frame information implantation recognition methods do not achieve the function of real-time or instant recognition of continuous frames. Users usually need to pause the video to the corresponding frame, select a frame or click to identify the target, and the feedback information is often directly superimpose...

AI Technical Summary

Problems solved by technology

[0003] Other existing continuous frame information implantation recognition methods do not achieve the function of real-time or instant recognition of continuous frames. Users usually need to pause the video to the corresponding frame, select a frame or click to identify the target, and the feedback information is often directly superimposed and displayed on the playing screen. The video affects the user's perception; there is also a lack of efficiency in processing frames. Because the total number of frames in a video is large, the process of terminal access to the server to obtain data takes a long time, and the video playback is in the deformation, scaling, and splicing of the scene. , Frequent movement, and handover will interfere with the analysis of video content, resulting in the speed and accuracy of information feedback falling short of requirements

Images