Research On Image Denoising Technology In The Face Recognition

Face recognition is one of the successful applications of image processing.It has received widespread attention in the past few decades.Although the existing face recognition systems have certain maturity in the theory of environment.But the objective conditions introduced by the environment,equipment,etc.still limit its successful application in actual.Due to the impact of illumination change in the environment and different equipment performance,the face image acquired by equipment always has some drawbacks such as containing noise and shadow,not enough contrast ratio.This paper focuses on the problem of noise in image,and proposed an adaptive nonlocal means denoising method.In addition,we combined the method with the characteristics of the face image structure and feasibility analysis,proposed a face recognition solution based on adaptive denoising method and implemented a local face recognition system on the Android platform.Nonlocal means is an effective denoising method,but the original nonlocal means may not find enough similar candidates for some non-repetitive image blocks.In order to mitigate these drawbacks,we propose an improved nonlocal means method using adaptive pre-classification in this paper.The proposed method employs Hu's moment and the threshold-based clustering algorithm to classify noisy image blocks adaptively.Then,a rotational block matching method is adopted to find the appropriate distance measurement between two blocks in an image.Experimental results shows that this method outperforms original nonlocal means both in quantitative and in visual qualities,especially when the image contains large amount of noise.According to the truth that the quality of denoised image is easily influenced by the filter parameter,we further improved the nonlocal means denoising method to make it better apply to face recognition solution.On the one hand,human face has symmetry structure characteristic and taking advantage of the seventh moment invariant could increase number of similar image block by recognizing mirror similar ones,which will contribute to achieve the purpose of improving the quality of image.On the other hand,an automatically setting scheme of filter parameter is applied for nonlocal filtering,which is applicable to the actual scene.This method can effectively reduce the effects of noise on face feature extraction,and contribute to the next step of face recognition.According to the proposed face recognition scheme,we implement a local face recognition system on the Android platform.The face recognition solution in the paper can effectively deals with the influence of noise on recognition,improves the face recognition rate and the user experience.