Illumination Normalization Method Research Articles

A Novel Approach of Face Recognition Using Optimized Adaptive Illumination–Normalization and KELM

Light variations from different directions on the face images cause severe performance degradation in face recognition system. These variations should be nullified or suppressed so that the recognition performance can be improved. Here, the objective is to develop a robust and pragmatic method which can compensate the effect of uncontrolled light incident from different directions on the person’s face. To normalize the effect of unpredictable illumination variations from face images, a novel illumination–normalization method is proposed by utilizing the fractional discrete cosine transform (Fr-DCT) and a nonlinear modifier. Fr-DCT provides a generalized α-domain in which illumination variation’s effect can be better processed. The proposed method adaptively selects and processes the low-frequency α-domain coefficients depending upon the light variations incident on the face image. The illumination-normalized images are classified with the help of a non-iterative classifier. The parameters of the proposed method are optimized with the help of genetic algorithm. The proposed method is tested over Extended YALE B, AR, CMU PIE and YALE face databases. Very promising results have been achieved with percentage error rate zero for CMU PIE and subset 3 of Extended YALE B, 0.38% for subset 4 and 0.55% for subset 5. On AR database, error rate is 0.5% when only illumination-varying images are considered for testing. The comparison of the results of the proposed method with those of the existing state-of-the-art methods clearly establishes the superiority of the proposed method.

ALCN: Adaptive Local Contrast Normalization

To make Robotics and Augmented Reality applications robust to illumination changes, the current trend is to train a Deep Network with training images captured under many different lighting conditions. Unfortunately, creating such a training set is a very unwieldy and complex task. We therefore propose a novel illumination normalization method that can easily be used for different problems with challenging illumination conditions. Our preliminary experiments show that among current normalization methods, the Difference-of-Gaussians method remains a very good baseline, and we introduce a novel illumination normalization model that generalizes it. Our key insight is then that the normalization parameters should depend on the input image, and we aim to train a Convolutional Neural Network to predict these parameters from the input image. This, however, cannot be done in a supervised manner, as the optimal parameters are not known a priori. We thus designed a method to train this network jointly with another network that aims to recognize objects under different illuminations: The latter network performs well when the former network predicts good values for the normalization parameters. We show that our method significantly outperforms standard normalization methods and would also be appear to be universal since it does not have to be re-trained for each new application. Our method improves the robustness to light changes of state-of-the-art 3D object detection and face recognition methods.

Normalization of face illumination with photorealistic texture via deep image prior synthesis

Illumination normalization for face recognition is of great importance when the face image is taken under a harsh lighting condition and only one single image is available. In this paper, an illumination normalization method that can preserve face identity and make the processed image have photorealistic face texture is proposed. First, an illumination regression filter is proposed to remove a large number of illumination components of the input image. Then, the accelerated proximal gradient algorithm is used for low-rank decomposition of the filtered image, which further reduces the residual illumination and noise of the face image. However, after these two operations, the face image has a peeled-off appearance. To avoid the unrealistic appearance, a deep image prior synthesis process is developed to synthesize the photorealistic face texture. To accomplish this task, the filtered image, the low-rank decomposed image, the original image and any a uniformly illuminated frontal face image as reference are used as synthetic ingredients. These ingredients, as prior knowledge, synthesize a face image that has the texture features of the reference image and retains the identity of the input image. We try to bridge the appearance gap between the synthetic ingredients and the final synthetic image by a deep neural network. Instead of training the network on a large dataset, the final synthetic image is generated by iterating the parameters of the generator-structured network according to a certain proportion of the synthetic ingredients. Systematic evaluations conducted on public databases demonstrate that the proposed method is robust to illumination with better performance than state-of-the-art methods.

An adaptive illumination normalization using non-linear regression for robust person identification under varying illuminations

The precision of person recognition using facial images in complex illumination conditions deteriorates severely. Various illumination normalization and compensation techniques alleviate the classification performance by linear and static modeling of illumination variations thus result in inappropriate illumination normalization. Here, the authors present novel and adaptive illumination normalization method with the help of kernel extreme learning machine (KELM). Few low frequency discrete-Cosine-transform (DCT) coefficients are used as input to KELM for learning the non-linear regression model for illumination variations. The trained KELM is used to find the illumination variations present in the input face image which is represented by the regression value. Depending upon the regression value corresponding to the illumination variations in the test image, the number of low frequency DCT coefficients is selected adaptively. The selected low frequency DCT coefficients are processed by a fuzzy filter. The processed face images are classified by KELM with polynomial kernel. The evaluation of the proposed illumination normalization method is conducted on Extended Yale B face database. Zero, 0.19 and 1.11 percent error rate on subset 3, 4 and 5, respectively, of this database proves the efficacy of the proposed method.

Single-sample face recognition under varying lighting conditions based on logarithmic total variation

The logarithmic total variation (LTV) algorithm is a classical algorithm that is proposed to address the illumination interference in face recognition. Some state-of-the-art techniques based on LTV assume that the illumination component mainly lies in the low-frequency features among face images. However, these techniques adopt unsuitable methods to process low-frequency features, resulting in final unsatisfactory recognition rates. In this paper, we propose an improved illumination normalization method based on the LTV method, called the RETINA&TH-LTV algorithm. In this algorithm, the retina model is utilized to eliminate most of the illumination component in low-frequency features. Then, an advanced contrast-limited adaptive histogram equalization technique is proposed to remove the residual lighting component. At the same time, through realizing threshold-value filtering on high-frequency features, the enhancement of facial features is achieved. Finally, the processed frequency features are combined to form a robust holistic feature image, which is then utilized for recognition. Insufficient training images in face recognition are also taken into consideration in this research. Comparative experiments for single-sample face recognition are conducted on YALE B, CMU PIE and our self-built driver databases. The nearest neighbor classifier and extended sparse representation classifier are employed as classification methods. The results indicate that the RETINA&TH-LTV algorithm has promising performance, especially in serious illumination and insufficient training sample conditions.

Adaptive illumination normalization via adaptive illumination preprocessing and modified weber-face

Illumination processing is a challenging task in face recognition. This paper proposes a novel illumination normalization method that aims to remove illumination boundaries and improve image quality under dark conditions. Firstly, to improve the image quality, an adaptive illumination preprocessing algorithm is adopted. Then we modify the Weber-Face model by suppressing the components which are greatly affected by the illumination. Experimental results on both Extended Yale B and CMU-PIE databases show that the proposed method can obtain high performance under complex illumination conditions. The accuracy on the Extended Yale B database is 93.02% and on the CMU-PIE database is 70.44%, which is the highest among the similar approaches. This method not only greatly improves the face recognition rate but also keep the computational complexity in low compared with several state-of-the-art methods.

Neuro-Fuzzy Hybridization using Modified S Membership Function and Kernel Extreme Learning Machine for Robust Face Recognition under Varying Illuminations

The multifaceted light varying environment severely degrades the performance of person recognition using facial images. Here, the authors present a novel person identification method using hybridization of artificial neural network (ANN) and fuzzy logic concepts. An efficient illumination normalization method is presented with the help of a new modified S membership function. The proposed method of illumination normalization retains the large scale facial features as well as suppresses the variations related to change in light variations. Kernel extreme learning machine (KELM) which is a nonlinear and non-iterative learning algorithm of ANN is used for classification. Various kernel types and parameters are experimented to find the best choice for robust classification. To assess the performance of proposed hybridization, Yale and extended Yale B face databases have been used. Very promising results have been achieved which establish the worth of the proposed method.

Face recognition under varying illumination based on singular value decomposition and retina modeling

Face recognition under the influence of complex illumination is a challenging problem to be solved. The common treatments for minimizing the affection of illumination variation are illumination preprocessing and illumination insensitive extraction techniques. However, the methods proposed previously present low performances. To realize high-accuracy recognition under varying illumination, this paper proposes a novel illumination processing algorithm called REC&SIG-SVD algorithm. Above all, singular value decomposition (SVD) is utilized to obtain preliminary high-frequency and low-frequency features of the face image in logarithm domain. This study proposes Sigmoid function which satisfies the principle of diminishing marginal utility to normalize singular values, aiming at calculating effective high-frequency features. Furthermore, this paper proposes a novel illumination normalization method to process low-frequency features, which is based on retina modeling cooperate with an advanced contrast limited adaptive histogram equalization (CLAHE). Meanwhile, enhancement on high-frequency features is realized by threshold-value filtering. Last but not least, the normalized high-frequency and enhanced low-frequency features are reassembled to form the normalized face image. The comparative trials based on Yale B and CMU PIE databases are conducted for our algorithm and other similar techniques as well as deep learning methods. The experimental results demonstrate that REC&SIG-SVD algorithm shows outstanding recognition performance.

Temporally Coherent Illumination Normalization for Indocyanine Green Video Angiography.

Indocyanine green (ICG) angiography is an imaging method for doctors to observe choroidal abnormalities in human eyes. The ICG angiograms typically exhibit inhomogeneous illumination, which poses serious difficulties for the development of computer-aided diagnostic tools. In this paper, we propose a novel illumination normalization method to alleviate the inhomogeneous illumination in ICG video angiograms. In particular, we first align the viewpoint of the input ICG video angiogram using an image registration method. Then, we acquire temporal information using time-dependent intrinsic image and compute the corresponding illumination image. Finally, we correct inhomogeneous illumination from the illumination image by estimating contrast and luminosity distortion. We have conducted extensive evaluation using ICG video angiograms of 60 patients. Two video quality assessment methods are utilized to evaluate the performance of our proposed illumination normalization method. The results show that our proposed method can help improve the visual quality of ICG video angiogram. Visual evaluation by a human expert also confirms that our method yields better illumination normalization results.

Illumination Robust Face Recognition Based on Image Fusion of Local Contrast Enhancement and Adaptive Smoothing

In this paper, a new face recognition method is proposed under variable illumination conditions. Firstly, the adaptive smoothing technique is improved according to a new conduction function and a parameter. And it can smooth face images well without enhancing edge effects. However, face images after smoothing are very white and their local contrast diminishes. Accordingly, we adopt the local contrast enhancement (LCE) to preprocess original images. And then, the improved adaptive smoothing (IAS) technology and LCE technology are used to fuse these images according to the technique based on orientation information measurement(OIM) in nonsubsampled Shearlet transform(NSST) domain. Experiments are done on the Yale B and CMU PIE face databases by using sparse representation based classification(SRC), and the results demonstrate that the proposed method has higher recognition rates in comparison to single preprocessing methods and some other illumination normalization methods, and it is robust even when the training set is single and under various lighting conditions.

Performance Analysis of Pixel based Face Recognition Methods

From the recent diversity of face recognition applications, it can be concluded that studies on face recognition are being actively conducted. However, high-performance algorithms sometimes cannot be adopted because of environmental reasons such as pose variation of targets or lighting conditions at the outdoor CCTV. Further, a low-resolution CCTV image cannot be used for an appearance-based high-performance method. Therefore, we consider pixel-based face recognition methods as CCTV applications. Before applying the methods, a Retinex-based illumination normalization method is used. The methods are applied with various facial poses, and three kinds of face recognition methods are comparatively analyzed in this paper. The experimental results showed that linear discriminant analysis exhibited the best performance as the matching method taking into consideration various facial poses.

Retinex-based illumination normalization using class-based illumination subspace for robust face recognition

Recent illumination normalization (IN) methods first decompose a face image into a reflectance (R)-image having a lighting-invariant characteristic and an illuminance (I)-image including shading and shadowing effects. An illumination-normalized I-image is then obtained by eliminating the lighting-dependent image variations (LDIV) from the I-image. Finally, the normalized I-and R-images are recombined for face recognition (FR). However, the decomposed-reflectance is often contaminated with the lighting effects. Moreover, the lighting normalization tends to remove the valuable discriminant information in the I-image. To address these problems, we employ the local edge-preserving filter to generate the R-image whereby the lighting-invariant information is well preserved. In addition, we propose a subspace-based IN method that can retain the large facial-structure in the I-image. To construct the proposed subspace, we calculate the LDIV within the same class of people from the training database of face images. Then, we apply the singular value decomposition to the calculated LDIV to obtain the basis images of the subspace. By projecting the I-image onto these basis images, we can effectively extract and eliminate the LDIV from the I-image without discarding the discriminant information. Experimental results confirm that FR with the proposed method outperforms that with existing IN methods under varying lighting conditions.

Enhanced facial texture illumination normalization for face recognition.

An uncontrolled lighting condition is one of the most critical challenges for practical face recognition applications. An enhanced facial texture illumination normalization method is put forward to resolve this challenge. An adaptive relighting algorithm is developed to improve the brightness uniformity of face images. Facial texture is extracted by using an illumination estimation difference algorithm. An anisotropic histogram-stretching algorithm is proposed to minimize the intraclass distance of facial skin and maximize the dynamic range of facial texture distribution. Compared with the existing methods, the proposed method can more effectively eliminate the redundant information of facial skin and illumination. Extensive experiments show that the proposed method has superior performance in normalizing illumination variation and enhancing facial texture features for illumination-insensitive face recognition.

A Robust Illumination Normalization Method Based on Mean Estimation for Face Recognition

An illumination normalization method for face recognition has been developed since it was difficult to control lighting conditions efficiently in the practical applications. Considering that the irradiation light is of little variation in a certain area, a mean estimation method is used to simulate the illumination component of a face image. Illumination component is removed by subtracting the mean estimation from the original image. In order to highlight face texture features and suppress the impact of adjacent domains, a ratio of the quotient image and its modulus mean value is obtained. The exponent result of the ratio is closely approximate to a relative reflection component. Since the gray value of facial organs is less than that of the facial skin, postprocessing is applied to the images in order to highlight facial texture for face recognition. Experiments show that the performance by using the proposed method is superior to that of state of the arts.

Shadow compensation and illumination normalization of face image

This study proposes a novel shadow compensation and illumination normalization method under uncontrolled light conditions. First, we decompose the face image into two images based on the Lambertian theory, which corresponds to the large- and small-scale features, respectively. Then, the threshold minimum-and-maximum filter on the small-scale features to smooth the shadow edge is applied. After that, the robust Principal Component Analysis and some normalization methods are used to remove the shadow and normalize the face image on the large-scale features. In the end, the normalized face image is obtained by combining both results from the large- and small-scale features. Our main contribution is that a more reliable shadow compensation approach is found, which can get a better normalized face image. Experiments on the Extended Yale B, CMU-PIE and FRGC 2.0 (Face Recognition Grand Challenge) face datasets show that not only the recognition performance is significantly improved, but also much better visual quality is achieved.

Illumination Invariant Face Recognition Using Nonlocal Total Variation in Logarithmic Domain

This paper presents a new illumination normalization method for robust face recognition under varying lighting conditions. In the proposed method, the illumination component is estimated by applying nonlocal total variation model in the logarithmic domain, and then the reflectance component is obtained based on reflectance model. The proposed method restrains the halo effect effectively while preserves the adequate texture information on the reflectance images. As an illumination invariant facial features, the reflectance images are directly utilized for face recognition. Experimental results on Yale face database B and CMU PIE database show that the performance of proposed method is robust and reliable in illumination invariant face recognition.

Face Recognition using Threshold Based DWT Feature Extraction and Selective Illumination Enhancement Technique

Face recognition (FR) under varying lighting conditions is challenging, and exacting illumination invariant features is an effective approach to solve this problem. In this paper, we propose a novel illumination normalization method called Selective Illumination Enhancement Technique (SIET) wherein the dark regions are selectively illuminated by using a correction factor which is determined by an Energy function. Also, we propose a Threshold based Discrete Wavelet Transform feature extraction for enhancing the performance of the FR system. Individual stages of the FR system are examined and an attempt is made to improve each stage. A Binary Particle Swarm Optimization (BPSO)-based feature selection algorithm is used to search the feature vector space for the optimal feature subset. Experimental results show the promising performance of Threshold based DWT extraction technique on ORL and UMIST databases and SIET on illumination variant databases like Extended Yale B and Color FERET.

Illumination Normalization using 2D Wavelet

Pose, expression and illumination variation are some challenges in face verification research area. When lighting condition changes, the appearance of a face varies intensely. In this paper, an illumination normalization method for face verification under various lighting conditions is presented. In the proposed approach, 2D wavelet is applied on the Yale database B and XM2VTS human face databases for illumination normalization in face verification. The performance of the method under illumination variations is improved compared to other pervious methods. Moreover, this procedure is very effective to compensate the shadows and also is very simple and it implements easily in real-time face verification system.

Comparative Analysis of various Illumination Normalization Techniques for Face Recognition

The change in facial appearance due to illumination variation degrades face recognition systems performance considerably. In this paper, various states of art illumination normalization techniques have been explained and compared. The classification of the image recognition has been done using artificial neural networks (ANN). We have compared four illumination normalization methods which are (1) discrete cosine transform (DCT) with rescaling of low frequency coefficients (2) discrete cosine transform (DCT) with discarding of low frequency coefficients (3) homomorphic filtering (HF) (4) gamma intensity correction (GIC). These methods are evaluated and compared on Yale and Yale B Faces databases.

스마트폰에서 웃음 치료를 위한 표정인식 애플리케이션 개발

본 논문에서는 스마트폰에서 웃음 치료를 위한 표정인식 애플리케이션을 제안한다. 제안된 방법에서는 스마트폰의 전면 카메라 영상으로부터 AdaBoost 얼굴 검출 알고리즘을 이용하여 얼굴을 검출한다. 얼굴을 검출한 다음에는 얼굴 영상으로부터 입술 영역을 검출한다. 그 다음 프레임부터는 얼굴을 검출하지 않고 이전 프레임에서 검출된 입술영역을 3단계 블록 매칭 기법을 이용하여 추적한다. 카메라와 얼굴 사이의 거리에 따라 입술 영역의 크기가 달라지므로, 입술 영역을 구한 다음에는 고정된 크기로 정규화한다. 그리고 주변 조명 상태에 따라 영상이 달라지므로, 본 논문에서는 히스토그램 매칭과 좌우대칭을 결합하는 조명 정규화 알고리즘을 이용하여 조명 보정 전처리를 함으로써 조명에 의한 영향을 줄일 수 있도록 하였다. 그 다음에는 검출된 입술 영상에 주성분 분석을 적용하여 특징 벡터를 추출하고 다층퍼셉트론 인공신경망을 이용하여 실시간으로 웃음 표정을 인식한다. 스마트폰을 이용하여 실험한 결과, 제안된 방법은 초당 16.7프레임을 처리할 수 있어서 실시간으로 동작 가능하였고 인식률 실험에서도 기존의 조명 정규화 방법보다 개선된 성능을 보였다. In this paper, we propose a recognition application of facial expression for laughter theraphy on smartphone. It detects face region by using AdaBoost face detection algorithm from the front camera image of a smartphone. After detecting the face image, it detects the lip region from the detected face image. From the next frame, it doesn't detect the face image but tracks the lip region which were detected in the previous frame by using the three step block matching algorithm. The size of the detected lip image varies according to the distance between camera and user. So, it scales the detected lip image with a fixed size. After that, it minimizes the effect of illumination variation by applying the bilateral symmetry and histogram matching illumination normalization. After that, it computes lip eigen vector by using PCA(Principal Component Analysis) and recognizes laughter expression by using a multilayer perceptron artificial network. The experiment results show that the proposed method could deal with 16.7 frame/s and the proposed illumination normalization method could reduce the variations of illumination better than the existing methods for better recognition performance.