Sketch Simplification Guided By Complex Agglomeration

Sketches are the prototype of product design.Designers don't need to pay attention to too much detail when drawing sketches.And they don't need to master the complicated tools in CAD as well.They only need paper and pen to complete the design quickly and smoothly.If you want to turn this native design into a vector graph that can be represented in a computer and can be edited further,you need to simplify the sketch and properly vectorize it.At present,this part of the conversion work sometimes needs to be done manually,that is,the sketch is imported into some editing software,and the manually trace the outlines.This work is time-consuming and laborious.In order to improve this working mode,the study of automatic sketch simplification algorithms has gradually become the focus of attention.Nowadays,vector drawing tools such as the tablet and the stylus are very popular.The sketch drawn by these tools is composed of a large number of strokes.The strokes are superimposed and overlapped with each other,and the spacing is not uniform,resulting in noisy outlines and obvious noise.This is one of the main difficulties in sketch simplification.Some of the popular simplification algorithms mainly rely on the idea of stroke grouping and merging to achieve sketch simplification.The problem is that the quality of the simplified results often depends on the setting of the thresholds.Secondly,some details of the image are ignored,and the completeness of the original image information cannot be guaranteed.Aiming at the analysis of the above problems,this article proposes a novel vector sketch simplification algorithm based on the basic complex.Combining the idea of an iterative contraction algorithm,the sketch is transformed into a representation similar to a two-dimensional planar mesh,i.e.,a base complex to explore a simplified approach.The base complex defined in this paper is a set consisting of different basic graphic primitives.First,the vector sketch is converted into a base complex.Then,under the constraint of the overall shape error minimization,the sketch is simplified by gradually agglomerate the base complex,that is,the simplification is realized by iteratively merging a pair of primitives(or eliminating a primitive),including curve-curve merging,patch-patch merging,and curve-patch merging.The shape difference measurement is used in the merging process to constrain the primitive pair merging order and control the degree of simplification.There are some non-smoothing situations exiting in the simplified base complex,such as the zigzag curve combination and the uncorrected intersections.This paper skillfully repairs the non-smoothing situations by the auxiliary information of the adjacent feature patches.Finally,the simplified base complex is converted into a single line vector graph represented by piecewise cubic Bezier curves to obtain the final simplified output.The simplified graph also implies the information of the feature patches in the original image.The degree of simplification of the sketch can be achieved by controlling the base complex using a a single thresholdIn this article we use the data sets provided from several current methods and many different styles of sketches for testing.It evaluates the algorithm from both efficiency and visual effects.The experimental results show that this algorithm can effectively reduce the complexity of the sketch,and the execution is efficient.Under the control of a single threshold we can get a good simplified result.When comparing with other methods,this paper show the advantages of retaining image detail information and integrity on subjective visual effects which are depicted in some pictures.In addition,we make a discussion on the setting of related parameters in the algorithm,algorithm limitations,and the future work.