Adaptive Context-Aware Filter Fusion for Face Recognition on Bad Illumination

Smart Library Automation System is increasingly attractive as an effective digital library system. Adapting automation in the library helps reduce the duplication of work, time-saving, and boosts work efficiency. One such significant feature is face recognition integrated. With this application, the system can use face recognition to enter and get the details of an end user. In recent years, face recognition has achieved many prodigious accomplishments based on Deep Convolutional Neural Networks (DCNN). In addition to constructing large face datasets, designing new effective DCNN architectures and loss functions are two trends to improve the performance of face recognition systems. Therefore, many studies have been published in state-of-the-art methods, making high-accuracy face recognition systems more possible than days in the past. However, it is still difficult for all research communities to train robust face recognition models because it depends heavily on their resources. This paper investigates and analyzes the effect of several effective loss functions based on softmax. Moreover, we also evaluate how hyper-parameter settings can impact the optimization process as well as the final recognition performance of the model trained by re-implementing these methods. The results of our experiments achieve state-of-the-art figures, which show the proposed method’s massive potential in improving face recognition performance. KeywordsFace recognitionDCNNEffective loss function

The performance of face recognition systems can be dramatically degraded when the pose of the probe face is different from the gallery face. In this paper, we present a pose robust face recognition model, centered on modeling how face patches change in appearance as the viewpoint varies. We present a novel model based on two robust local appearance descriptors, Gabor wavelets and Local Binary Patterns (LBP). These two descriptors have been widely exploited for face recognition and different strategies for combining them have been investigated. However, to the best of our knowledge, all existing combination methods are designed for frontal face recognition. We introduce a local statistical framework for face recognition across pose variations, given only one frontal reference image. The method is evaluated on the Feret pose dataset and experimental results show that we achieve very high recognition rates over the wide range of pose variations presented in this challenging dataset.

Illumination, pose variations, disguises, aging effects and expression variations are some of the key factors that affect the performance of face recognition systems. Face recognition systems have always been studied from a recognition perspective. Our emphasis is on deriving a measure of similarity between faces. The similarity measure provides insights into the role each of the above mentioned variations play in affecting the performance of face recognition systems. In the process of computing the similarity measure between faces, we suggest a framework to compensate for pose variations and introduce the notion of 'half-faces' to circumvent the problem of non-uniform illumination. We used the similarity measure to retrieve similar faces from a database containing multiple images of individuals. Moreover, we devised experiments to study the effect age plays in affecting facial similarity. In conclusion, the similarity measure helps in studying the significance facial features play in affecting the performance of face recognition systems.

In this paper, we propose a novel robust face recognition framework named nuclear norm based superposed collaborative representation classifier (NNSCRC) to handle illumination variations, occlusion and undersampled problems in face recognition. Specifically, we develop a superposed linear collaborative representation classifier for robust face recognition by representing the query image in terms of a superposition of the class centroid, the shared intra-class difference, and the low rank error. By representing a face image as the class centroid and the shared intra-class difference, our model can effectively enhance the face recognition performance on undersampled databases. In addition, since the occlusion and illumination variations generally lead to a low-rank error image, we use nuclear norm matrix regression to obtain these low-rank errors, which makes our model able to reconstruct the test image better. Extensive experiments are performed on Extended Yale-B and AR databases, which show the effectiveness of NNSCRC in robust face recognition.

Face aging simulation and recognition based on NMF algorithm with sparseness constraints

The challenges imposed by the Single Sample Per Person face recognition problem become especially critical when face images are captured in uncontrolled environments, which typically involve variations in illumination, facial expression, pose, occlusion, and more. In particular, illumination changes caused by poor lighting conditions can significantly degrade the performance of face recognition systems. In this paper, we investigate the SSPP face recognition problem in the presence of potential illumination variation and propose a two-fold approach. Firstly, we analyse and quantify the illumination variations in face images and then normalise these variations based on a probabilistic image enhancement approach. Subsequently, the feature embedding process was performed using the deep learning-based FaceNet system, and finally, faces were classified using the classical Support Vector Machine. The performance of the proposed approach is demonstrated through comprehensive experiments compared to state-of-the-art techniques. The proposed method achieved significant accuracy improvements, attaining 96.17% and 95.68% for 20 and 30 subjects, respectively, on the Extended Yale-B dataset. The performance of the proposed method is evaluated and exhibits superiority in handling illumination compared to state-of-the-art counterparts.

Illumination variations significantly affect the performance of the automatic face recognition system. To achieve optimum contrast enhancement, contrast limiting adaptive histogram equalization (CLAHE) has been used in this work. Histogram Equalization (HE) modifies the histogram of the image globally based on intensity distribution of an entire image. However, the feature of interest in an image needs enhancement locally. CLAHE is based on the intensity distribution in a neighborhood of every pixel in the image. Further, for removing the illumination variations in the face image, the appropriate number of low frequency DCT coefficients has been scaled down as illumination variations mainly lie in the low-frequency band. After eliminating illumination variations effect, mapping of the data on to another feature space is done using Kernel PCA (KPCA), which extract higher order statistics. KPCA has the advantage of less computation time and improve performance level. Classification is done by using nearest neighbor classifier. Experiments are performed on Extended Yale B database. The experimental results show that the performance of our method is significantly better than that of any existing state-of-art technique.

Robust face recognition using 2D and 3D data: Pose and illumination compensation

In this paper evolutionary classifier fusion method is used to optimize the performance of face recognition system. Initially different illumination environments are modeled as multiple contexts using unsupervised learning and then optimized classifier ensemble are searched for each context using genetic algorithm (GA). For each context multiple optimized classifiers are searched each of which are referred as context based classifier. Then evolutionary framework of combination of such classifiers is applied to optimize the face recognition as a whole. Evolutionary classifier fusion is compared with the single classifier system. Experiment is done using real time Inha database and FERET database. Experimental results show that the proposed multiple context based fusion method gives superior performance than the method without using fusion and optimize face recognition performance.

In recent days, there has been increasing need for recognition of unconstrained face images, such as those collected from the web or captured by mobile devices and video surveillance cameras. In such real world scenarios, human faces could be easily occluded by other objects that make the face recognition task as a complex one. Satisfactory performance has been achieved earlier but often only in the controlled environments. But it is very tedious to obtain holistic face images for unconstrained face recognition. Thus, in order to avoid the degradation of face images and the huge variations due to illumination, pose, occlusion and expression, a new Robust Face Recognition approach is proposed. In this approach, the partial face recognition using Scale Invariant Feature Transform (SIFT) technique is combined with Multi-directional Multi-level Dual Cross Patterns (DCP) technique that makes the recognition task as a robust one when compared to other face recognition approaches. Then, the Robust Point Set Matching (RPSM) is used to match the corresponding stable keypoints from both the gallery image and probe face image. Finally, PNN and K-NN classification are used to classify the face images even with the presence of occlusion, random partial crop, illumination, pose and exaggerated facial expression. The proposed robust face recognition system is evaluated based on the performance parameters such as sensitivity, specificity, accuracy, precision and recall.

The performance of face recognition system degrades when the variability of the acquired faces increases. Prior work alleviates this issue by either monitoring the face quality in pre-processing or predicting the data uncertainty along with the face feature. This paper proposes MagFace, a category of losses that learn a universal feature embedding whose magnitude can measure the quality of the given face. Under the new loss, it can be proven that the magnitude of the feature embedding monotonically increases if the subject is more likely to be recognized. In addition, Mag-Face introduces an adaptive mechanism to learn a well-structured within-class feature distributions by pulling easy samples to class centers while pushing hard samples away. This prevents models from overfitting on noisy low-quality samples and improves face recognition in the wild. Extensive experiments conducted on face recognition, quality assessments as well as clustering demonstrate its superiority over state-of-the-arts. The code is available at https://github.com/IrvingMeng/MagFace.

Prior research has shown that the performance of face recognition systems in variant expression condition degrade seriously compared with invariant expression condition. Face recognition with variant expression is a challenging problem. In this paper, SIFT method is proposed to research on face recognition with variant expression. Two experiments using SIFT method are performed on a variant expression face database. In experiment 1, two images of one person with different expression are compared. In experiment 2, two images of different persons with the same expression are compared. Experimental results show that SIFT method could overcome the whole comparability of different faces and extract the local detail features of face. The experiment demostrates the huge potential of SIFT method in application to face recognition with variant expression.

In this paper, face recognition algorithms under three low-quality conditions, namely, complex illumination, low resolution and partial occlusion, are summarized, and relatively new face recognition algorithms under low-quality conditions are introduced and explained. Such as the algorithm of face recognition with occlusion based on partition, face recognition algorithm based on Iteratively Reweighted Robust Principal Component and face recognition with occlusion based on MEBML, illumination face recognition based on RPCA and Convolutional Neural Network, face recognition based on Illumination Normalization and Block-based Adaptive Local Ternary Pattern, low-resolution face recognition based on blocking CS-LBP and weighted PCA algorithm, pose Robust low-resolution face recognition via CKEDA and other face recognition algorithm under low-quality conditions. The existing problems and development trend of face recognition under low quality conditions are analyzed.

Joint representation and pattern learning for robust face recognition

Pose variation which brings illumination change, occlusion and non-linear scale variation, dramatically drops the performance of face recognition systems. In this paper, we propose a novel pose invariant face recognition method, in which we build a joint sparse coding scheme to predict face images from a certain pose to another. By introducing autoregressive regularization and symmetric information, our algorithm could achieve high robustness to local misalignment and large pose differences. Besides, we propose a new coarse pose estimation algorithm by collaborative representation classifier, which is very fast and enough accurate for our synthesis algorithm. The experiment results on multi-pose subsets of CMU-PIE and FERET database show the efficiency of the proposed method on multi-pose face recognition.