Font Size: a A A

Hand-drawn Sketches Based Image Retrieval Technology And Algorithm Implementation

Posted on:2018-01-07Degree:MasterType:Thesis
Country:ChinaCandidate:Q Y WuFull Text:PDF
GTID:2348330518970043Subject:Electronic Science and Technology
Abstract/Summary:PDF Full Text Request
With the expansion of the image in the digital image library and the popularity of high-tech hardware such as touch-screen mobile phone and tablet computer,the hand-drawn sketch retrieval technology has more and more influence on people's lives.In recent years,the image retrieval technology has drawn a lot of attention from domestic and foreign experts who have made a lot of contributions.Image retrieval technology has experienced a leap-forward development from text-based search to image-based keyword search,and hand-drawn sketch retrieval search.In the sketch retrieval system,people can draw the target image on the drawing board provided by the system,and then the system will process the edge extraction,contour tracking,feature point assignment,feature conversion,similarity matching and so on,which provides a number of options for the purpose of the image.In the process of searching the target image,the users mainly depend on the shape information of the image,which is also in line with the characteristics of the human eye observation.Therefore,the sketch retrieval system only considers the outline of the image characteristics regardless of the characteristics of color and texture features.Based on the research of advanced sketch retrieval technology at home and abroad,this paper improves the algorithm of sketch retrieval and designs a new image retrieval system.According to the features of drawing sketch from users,we add the sketch preprocessing module.In addition,this paper compares the performance of different edge extraction algorithms,and proposes three-neighborhood tracking method of x-axis and y-axis direction by optimizing the five-neighborhood tracking method proposed by Wang Fusheng,which not only improving the tracking accuracy of the contour points but also reducing the time consumption of the system.Then,paper improve the shape context algorithm by adopting the Fourier transform so that the improved shape context algorithm has a rotation invariance which solving the problem that the image offset will reduce search accuracy.Finally,this paper designs ahand-drawn sketch retrieval system according to the improved algorithm,which is experimented with a small library with 300 images,and has achieved good results.
Keywords/Search Tags:Sketch retrieve, Feature extraction, Shape context, Canny edge detection, Fourier transform
PDF Full Text Request
Related items