The Research Of Object Recognition Technology Based On Improved BING Algorithm

Object recognition, as one of the research direction of computer vision, the maingoal is to identify the object in the image or video based on pattern recognitionmethods, which describe the locations, shape and direction information of objects.The current object recognition technology has very broad application and demandscenarios in the military field, civil life field and office work, such as computerforensics, ID identification, face recognition, intelligent access control system, bodyfeeling game and medical image. This paper will develop object recognitiontechnology based on the BING algorithm in-depth.This paper proposes three algorithms for object recognition, respectively,e-GSNG algorithm, nu-GSNG algorithm and BING-Frac algorithm. The e-GSNGalgorithm and the nu-GSNG algorithm is the improved algorithm with the fusion ofthe NG algorithm and the LIBSVM library which replace the LIBLINEAR library.BING-Frac algorithm was proposed based on fusion of BING algorithm and thefractal technology, with the local fractal dimension as NG computing feature inBING-Frac algorithm.This paper uses the simulation experiments validated for the three objectrecognition algorithm. The test results confirmed that the three algorithms can havecertain advantages in some aspects. BING-Frac algorithm can be compared to thecurrent frontier BING algorithm with considerable efficiency, time and memory, andcan have more excellent DR recognition efficiency in the lower candidate windownumber. The performance advantages of both the e-GSNG algorithm and nu-GSNGalgorithm are similar, and both can gain higher DR recognition efficiency at theexpense of running time.After the above-mentioned theoretical research, this paper develops mobile application practice based on BING-Frac algorithm. The goal of the application is todevelop object recognition applications based on Android software platform, whichhas the advantages of high efficiency and real-time. Mobile application frameworkincludes Android client part and the web server part, which respectively play the roleof image extraction and transmission, image process and return.