Research On Cartoon Copyright Protection Technology Based On Character’s Individual Action Characteristics

With the growth and popularization of network technology and the sharp increase of the number of internet users,the cartoon industry is facing new opportunities.However,the continuous emergence of pirated and plagiarized cartoons restricts the development of the cartoon industry.Modern comics have been transformed from traditional paper comics to digital comics,whose fast propagation speed and wide dissemination provide the prerequisite for the appearance of pirated and plagiarized comics.At the same time,the degree of legal protection is not enough,tort determination is difficult.Therefore,positive measures should be taken to improve the relevant technology of cartoon copyright protection to prevent the infringement of cartoon copyright fundamentally.The earliest meaning of cartoon copyright is the version and the right,that is,the right to copy.But with the changes in the way comics are distributed,it is becoming easier to pirate and copy,and the infringement behavior of infringers is becoming more and more intense.The criterion to judge whether the infringement is caused is to imitate the expression of the original work,while copyright protection protects the creative expression of the author,that is,whether the unique expression characteristics are used.As an important feature to show the identity,inner activity and ideological quality of the characters,the character’s "individual action" in the cartoon is particularly important to the characters.Therefore,this paper proposes a method of cartoon copyright protection based on character’s individual action characteristics.This method can be used to compare the characteristics of original cartoon actions with those of pirated or plagiarized actions to identify the existence of pirated or plagiarized actions.Firstly,an image preprocessing scheme for character extraction is designed.As most modern comics are color comics with high color saturation,obvious color contrast and great variation of image boundary,Grab Cut algorithm with good segmentation effect in the field of color image segmentation is selected for the selection of cartoon character extraction method.In the implementation of the algorithm,the original automatic segmentation rectangle of Grab Cut algorithm is abandoned,and interactive foreground segmentation is used instead to better obtain cartoon characters.After that,binarization and morphological processing are carried out on the segmentation results of cartoon characters to eliminate the influence of multi-connected domains brought by interference points and lines,making preparation for the feature extraction of cartoon characters’ personality actions.Secondly,a feature extraction scheme based on character profile,moving character and character posture is designed.By extracting the perimeter and area of the character profile,the aspect ratio and tilt angle of the smallest rectangle are added to get the character profile feature.The three-dimensional coordinate is established by the change of the role in the three-dimensional space,and the motion angle characteristics of the role are calculated.The position of the central point of the role in the two-dimensional image is the first two dimensions of the three-dimensional space,and the change of the zoom in and zoom out of the role is the third dimension of the three-dimensional space.Considering the body proportion of cartoon characters and the relative randomness of drawing technique,the pose characteristics of the characters were extracted by skeletonization,and the redundant branches are removed by deburring operation.Then the joint angle information of the skeleton is obtained by the method of obtaining the joint angle of the skeleton.Finally,through the recognition analysis results of character contour feature,moving character feature and character pose feature,it is found that only using a certain feature has limitations in character action recognition.In order to improve the accuracy of action recognition,a weighted fusion of the three methods is proposed,and a threshold value is set to judge whether there is piracy or plagiarism in character actions to achieve the purpose of copyright protection.The experimental results show that the proposed method of cartoon copyright protection based on character’s individual action features has a good effect on the recognition of character’s individual action,and can accurately extract character’s contour features,moving character features and character posture features.The method in this paper breaks the limitation of the original single-angle feature recognition,and can recognize the cartoon character actions more comprehensively through multi-angle feature extraction,which has good practicability.