Font Size: a A A

Partly Area Of Image Low-poly Style Conversion

Posted on:2020-10-16Degree:MasterType:Thesis
Country:ChinaCandidate:Y W L OuFull Text:PDF
GTID:2415330611982444Subject:Computer technology
Abstract/Summary:PDF Full Text Request
"Low-poly" is an emerging art style.The style is characterized by the use of a large number of triangles of different shapes and sizes to represent the image.The near-looking or magnifying view is a triangular color block,but it can be viewed or reduced.See the overall shape of the object.There have been some algorithms for converting the style of the entire image to the Low-poly style,but there has been no research on the Low-poly style conversion for the local area of the image.Our method proposes a Low-poly stylization method for only partly regions of images.Firstly,the image segmentation algorithm is used to obtain the Low-poly stylized region by the user's small interaction.Then,two edge detection algorithms are combined to detect the complete edge line of the image,and the adaptive seed point selection is used to select different regions of the image.Seed points and generate triangles and fill the corresponding colors through the seed points in the specified area.From a geometric point of view,the use of finite triangles is not able to completely reconstruct the irregular image of the edge.These regions that cannot be reconstructed by the triangle greatly affect the final style conversion effect,so the problem that the triangle cannot completelycover the selected region is used.The Low-poly style conversion is completed by the improved image restoration algorithm to repair the area in the specified area where the triangle cannot be completely covered by the Low-poly stylization.The method can get better Low-poly stylization effect on the local area of the image.The different style areas of the image are clearly distinguished,the image details remain relatively complete,and the great Low-poly art style is displayed after the repair algorithm repairs the residual area.
Keywords/Search Tags:Low-poly, edge detection, mesh segmentation, image segmentation, image restoration
PDF Full Text Request
Related items