CAJ | 학술논문

现阶段基于单一的特征提取算法已不能满足人们对图像查准率、查全率的要求.有效地整合图像的颜色、纹理、形状等特征,使得各特征优势互补,进而提高系统检索效率现已成为一个值得研究并深入的课题[1].从颜色、纹理两方面的特征出发,研究并设计了一整套的方案,很好的解决了查准率、查全率等问题.
현계단기우단일적특정제취산법이불능만족인문대도상사준솔、사전솔적요구.유효지정합도상적안색、문리、형상등특정,사득각특정우세호보,진이제고계통검색효솔현이성위일개치득연구병심입적과제[1].종안색、문리량방면적특정출발,연구병설계료일정투적방안,흔호적해결료사준솔、사전솔등문제.
At present,based on single feature extraction algorithm already cannot satisfy the requirements of precision rate and recall rate. The effective integration of image color, texture, shape characteristics, have complementary advantages, and improve the retrieval efficiency system that becomes a worthy of study and in-depth topic now. From color, texture characteristics,this paper researchs and designs a set of solutions that solve the precision rate and recall rate problems.