Face Detection Algorithm Research Articles

Facial expression recognition based on eigenspaces and principle component analysis

Facial expression detection or emotion recognition is one of the rising fields of research on intelligent systems. Emotion plays a significant role in non-verbal communication. An efficient face and facial feature detection algorithms plays as important role in indentifying of an emotion of a person at a particular moment. In this work, the authors implemented a system that recognises a person's facial expressions from the input images, using the algorithm of eigenspaces and principle component analysis (PCA). Eigenspaces are the face images which are projected onto a feature space that encodes the variation among known face images. PCA is used in this paper to make dimensional reduction of images in order to obtain a reduced representation of face images. The implementation is applied on three different facial expressions databases, extended Cohn-Kanade facial expression database, Japanese female facial expression database and self-made database in order to find out the effectiveness of the proposed method.

Facial expression recognition based on eigenspaces and principle component analysis

Facial expression detection or emotion recognition is one of the rising fields of research on intelligent systems. Emotion plays a significant role in non-verbal communication. An efficient face and facial feature detection algorithms plays as important role in indentifying of an emotion of a person at a particular moment. In this work, the authors implemented a system that recognises a person's facial expressions from the input images, using the algorithm of eigenspaces and principle component analysis (PCA). Eigenspaces are the face images which are projected onto a feature space that encodes the variation among known face images. PCA is used in this paper to make dimensional reduction of images in order to obtain a reduced representation of face images. The implementation is applied on three different facial expressions databases, extended Cohn-Kanade facial expression database, Japanese female facial expression database and self-made database in order to find out the effectiveness of the proposed method.

Robotic Arm-Based Face Recognition Software Test Automation

Facial recognition is a feature that uses facial detection algorithms to detect a face and then invokes facial recognition algorithms to try to match the person’s face. First, a person’s face needs to be enrolled; during enrollment, the person’s facial details are saved in a database—in this case, a mobile phone. The facial recognition algorithm uses this database to match the currently presented face with the faces saved in the database. The efficiency of a facial recognition algorithm depends upon the speed at which it can detect and recognize faces. The problem we address in this paper is that a reliable, automated method for testing facial recognition features in mobile phones. Because the multimedia capabilities of smartphones have expanded phenomenally, the need for thoroughly testing facial recognition algorithms has become crucial. The uses of facial recognition can range from facial authentication for unlocking phones to security for a number of other applications. To meet these needs, a reliable, automated test for validating the facial recognition algorithms must be developed. The challenge for the software test team was to automate the test cases, which involve tilting the phone at specific angles from the test subject. The permissible angular movements of the phone are determined by the algorithmic specifications of the facial recognition algorithm. We tested scenarios involving multiple faces, motion blur, panning the phone in front of the test subject faces at various speeds, and so on. We adopted a robotic arm to perform the facial recognition test cases and developed software to program the robotic arm to test the phone’s facial recognition software for functionality, performance, and stability as well as adversarial tests. This paper discusses the computer vision use case for facial recognition and describes how we developed an automated facial recognition test suite using a robotic arm.

HyperFace: A Deep Multi-Task Learning Framework for Face Detection, Landmark Localization, Pose Estimation, and Gender Recognition.

We present an algorithm for simultaneous face detection, landmarks localization, pose estimation and gender recognition using deep convolutional neural networks (CNN). The proposed method called, HyperFace, fuses the intermediate layers of a deep CNN using a separate CNN followed by a multi-task learning algorithm that operates on the fused features. It exploits the synergy among the tasks which boosts up their individual performances. Additionally, we propose two variants of HyperFace: (1)HyperFace-ResNet that builds on the ResNet-101 model and achieves significant improvement in performance, and (2)Fast-HyperFace that uses a high recall fast face detector for generating region proposals to improve the speed of the algorithm. Extensive experiments show that the proposed models are able to capture both global and local information in faces and performs significantly better than many competitive algorithms for each of these four tasks.

Pre-detection Technology of Clothing Image Segmentation Based on GrabCut Algorithm

Image segmentation, as a key technology in digital image processing, serves as the basis for image processing and analysis and understanding. Its main purpose is to separate useful image information (i.e. foreground) from useless image information (i.e. background) in some way. With segmentation of clothing image in online shopping as the center, this paper studied image segmentation technology. Study and analysis of a large number of online clothing images found that image clothing can be roughly divided into two categories. The first is simply displayed clothing without human model, with several clothing displayed together. The other is image with human model. By distinguishing the two types of images with face detection algorithm and edge detection method, applying different location algorithms for the two types of images, and adjusting image location with iterative algorithm, ultimately, more accurate localization frames were obtained, which can replace the part in traditional classical GrabCut algorithm that requires manual participation, and realize automatic batch operation of image segmentation. The final test data proved effectiveness of the new improved algorithm which can be applied in retrieval system of mass images at the time of online clothing shopping.

Adaptive Kalman Filter Applied to Vision Based Head Gesture Tracking for Playing Video Games

This paper proposes an adaptive Kalman filter (AKF) to improve the performance of a vision-based human machine interface (HMI) applied to a video game. The HMI identifies head gestures and decodes them into corresponding commands. Face detection and feature tracking algorithms are used to detect optical flow produced by head gestures. Such approaches often fail due to changes in head posture, occlusion and varying illumination. The adaptive Kalman filter is applied to estimate motion information and reduce the effect of missing frames in a real-time application. Failure in head gesture tracking eventually leads to malfunctioning game control, reducing the scores achieved, so the performance of the proposed vision-based HMI is examined using a game scoring mechanism. The experimental results show that the proposed interface has a good response time, and the adaptive Kalman filter improves the game scores by ten percent.

Applications and techniques in information and network security

Applications and techniques in information and network security

A parallel approach for the training stage of the Viola-Jones face detection algorithm

A parallel approach for the training stage of the Viola-Jones face detection algorithm

Non-contact remote estimation of cardiovascular parameters

Abstract Cardiovascular disease is a serious threat to human health. It is crucial to monitor the cardiovascular parameters reliably and conveniently. Non-contact measurement has been widely studied. However, there are some inevitable factors that limit the use of the platform and even lead to inaccuracy estimation. Hence, a novel non-contact method that estimates cardiovascular parameters under ambient light is proposed. The most suitable region of interest (ROI) is determined by a colormap, which is a map consists of Fast Fourier Transform (FFT) peak amplitude of every pixel. Comparisons suggest that the region includes cheeks and nose is the most appropriate, followed by the forehead. To remove the motion artifacts caused by body movement, face detection and tracking algorithms are performed. Further, a multi-step signal process that combines independent component analysis (ICA) and wavelet de-noising is utilized to extract clean signal from the noise-corrupted raw waveform. Moreover, various distance between subject and camera and the change of ambient light intensity are considered, where statistics results reveal that this non-contact methodology is blind to these factors. Compared with the gold standard pulse oximeter, the proposed method shows a high accuracy even in the presence of motion artifacts.

Development of a Face Detection Algorithm Based on Skin Segmentation and Facial Feature Extraction

This paper presents a face detection algorithm capable of detecting face(s) without prior training as a face classifier. The technique employed in developing the algorithm is based on skin segmentation and facial feature extraction of the two eyes and mouth. Skin segmentation was done in the red, green, blue color space. White balance correction was employed to correct the change in image temperature that occurs due to change in lighting conditions at the point of acquiring image. Morphological operations and bounding box were employed to search and extract face region from the segmented skin region. Facial feature, eyes and mouth, were extracted for final verification of the sensed face using the Laplacian of Gaussian filter and the isosceles triangle matching rules. The extracted features were used to develop the face detection algorithm. The developed algorithm was evaluated using random images taken under different lighting conditions. Furthermore, the efficiency of the developed face detection algorithm was evaluated using a standard face detection image database. The result obtained shows that the developed face detection algorithm performed satisfactorily well with 81.37% detection accuracy. Furthermore, the results obtained from the performance evaluation of the developed face detection for this study has shown it clearly that accuracy detection of dissimilar faces in images with complex background is possible and attainable

Analysis of methods and technologies of human face recognition

The object of research is the processes of biometric identification and human authentication based on the image of his face for computer vision systems. One of the most problematic places in biometric identification systems using computer vision is the problem of eliminating ambiguity of «scanning». Such ambiguity arises when designing three-dimensional objects of the real world on flat images. In the course of the research, the results of the analysis of the effects of requirements and factors on the features and characteristics of the object of the biometric face recognition system are used. First of all, it is the variability of visual images, the design of three-dimensional objects, the number and location of light sources, the color and intensity of radiation, shadows or reflections from surrounding objects. The solution to the problem of detecting objects on the image lies in the correct choice of the description of objects, for the detection and recognition of which the system is created. Analysis of the features of classes and the properties of face recognition tasks shows that it is sufficient for a database of authentication systems to store a small set of predefined key characteristics, as much as possible characterize the images. Thus, by configuring the system to reduce the probability of incorrect identification, it is possible to use several images belonging to one person. For such purposes, a video sequence of certain specific head movements and facial muscles of the face is sufficient. A generalized algorithm for automatic face detection and recognition is developed. The presented scheme of the generalized algorithm consists of nine simple steps and takes into account the identification features using photo and video images. The advantage of the algorithm is the simplicity of implementation, it allows already at the design stage of the identification system, to quickly evaluate the system's operability by analyzing the internal interaction of its elements.

Face Detection Based on Skin Color and Adaboost Algorithm

The problem of multiple face detection and missed detection caused by illumination and complex background is avoided. A multi face detection algorithm based on skin color and Adaboost algorithm is proposed. The first chapter of this article is to apply the CbCrCg space based on the least squares of the skin color clustering model as the pretreatment process of face detection, And then you need to go further and take the area as an input image and you go through the Adaboost cascade classifier for face detection, To further remove non-face areas in the skin area, Finally, a more accurate face orientation is achieved.

On the Extraction of Anthropometric Parameters by Visual and Non-Visual Means

In this paper the system for collection and recording of anthropometric data is presented, along with the novel techniques for extraction of such data. The very system is built on selected open-source platform having developed various plugins as a part of the project. Means for the extraction follow two approaches: visual and non-visual. The first presumes the acquiring of data from static 2D image, the latter gets data through the direct measurement. Visual approach utilizes several principles following the image processing and related face detection algorithms. Moreover, known anthropometric relations are utilized to estimate other human body proportions. As for the non-visual approach, the hardware for direct measurement of human body parameters is designed, implemented and tested in the real environment. The output in the form of data of individual user may serve for statistical comparison with other users. Further, data of all users is to be used for correlation studies with several diseases and changes of overall health condition. In addition, the extracted data follow the concept of newly proposed Human body description language (HBDL) that may be used in various scientific fields and applications or various programming languages. Thus providing the standardized and structured form of data entry.

A Functional Regression Approach to Facial Landmark Tracking.

Linear regression is a fundamental building block in many face detection and tracking algorithms, typically used to predict shape displacements from image features through a linear mapping. This paper presents a Functional Regression solution to the least squares problem, which we coin Continuous Regression, resulting in the first real-time incremental face tracker. Contrary to prior work in Functional Regression, in which B-splines or Fourier series were used, we propose to approximate the input space by its first-order Taylor expansion, yielding a closed-form solution for the continuous domain of displacements. We then extend the continuous least squares problem to correlated variables, and demonstrate the generalisation of our approach. We incorporate Continuous Regression into the cascaded regression framework, and show its computational benefits for both training and testing. We then present a fast approach for incremental learning within Cascaded Continuous Regression, coined iCCR, and show that its complexity allows real-time face tracking, being 20 times faster than the state of the art. To the best of our knowledge, this is the first incremental face tracker that is shown to operate in real-time. We show that iCCR achieves state-of-the-art performance on the 300-VW dataset, the most recent, large-scale benchmark for face tracking.

Face detection of golden monkeys via regional color quantization and incremental self-paced curriculum learning

Animal detection plays a very vital role in wildlife protection and many other real life applications. In this paper, we focus on face detection of Golden monkeys who live in Qinling Mountains, Shaanxi province, China, and present a relatively complete face detection algorithm to detect these monkeys’ faces, which mainly includes three parts: the location of the monkeys’ bodies, the detection of th+e suspicious facial skin and the accurate detection of the true faces. Firstly, regional color quantization is proposed to quantize the HSV color space for the nature images with different sizes, and we can get the areas of the monkeys’ bodies according to the color distribution of the monkeys’ hair in the histogram of the quantized color space. Then the areas of suspicious facial skin can be extracted from these areas of the monkeys’ bodies. Further, we propose incremental self-paced curriculum learning (ISPCL) to detect the true monkeys’ faces accurately. In our method, regional color quantization can increase the color differences between the background and the monkeys’ hair, so that the segmented results have fewer background pixels. Besides, the basic idea of the incremental learning is introduced into the training process of SPCL, which is to simulate the process in which human learn something from easy samples to hard ones, this idea is able to improve the performances of face detection. The experimental results demonstrate that the proposed algorithm can locate the monkeys’ bodies in the images with different sizes, and can detect their faces effectively. This research lays a foundation for face recognition and behavior analysis of golden monkeys.

Semi-supervised annotation of faces in image collection

The objective of this work is to correctly detect and recognize faces in an image collection using a database of known faces. This has applications in photo-tagging, video indexing, surveillance and recognition in wearable computers. We propose a two-stage approach for both detection and recognition tasks. In the first stage, we generate a seed set from the given image collection using off-the-shelf face detection and recognition algorithms. In the second stage, the obtained seed set is used to improve the performance of these algorithms by adapting them to the domain at hand. We propose an exemplar-based semi-supervised framework for improving the detections. For recognition of images, we use sparse representation classifier and generate seed images based on a confidence measure. The labels of the seed set are then propagated to other faces using label propagation framework by imposing appropriate constraints. Unlike traditional approaches, our approach exploits the similarities among the faces in collection to obtain improved performance. We conduct extensive experiments on two real-world photo-album and video collections. Our approach consistently provides an improvement of $${\sim } 4$$ % for detection and $$5{-}9$$ % for recognition on all these datasets.

Deep learning for content-based video retrieval in film and television production

While digitization has changed the workflow of professional media production, the content-based labeling of image sequences and video footage, necessary for all subsequent stages of film and television production, archival or marketing is typically still performed manually and thus quite time-consuming. In this paper, we present deep learning approaches to support professional media production. In particular, novel algorithms for visual concept detection, similarity search, face detection, face recognition and face clustering are combined in a multimedia tool for effective video inspection and retrieval. The analysis algorithms for concept detection and similarity search are combined in a multi-task learning approach to share network weights, saving almost half of the computation time. Furthermore, a new visual concept lexicon tailored to fast video retrieval for media production and novel visualization components are introduced. Experimental results show the quality of the proposed approaches. For example, concept detection achieves a mean average precision of approximately 90% on the top-100 video shots, and face recognition clearly outperforms the baseline on the public Movie Trailers Face Dataset.

Abstract LB-271: SplashRNA, a sequential classification algorithm for ultra-potent RNAi

Abstract We present SplashRNA, a sequential classifier - analogous to face detection algorithms - to predict ultra-potent microRNA-based short hairpin RNAs (shRNAs) for virtually any gene. Trained on existing and novel large-scale datasets, SplashRNA outperforms previous algorithms and reliably predicts the most efficient shRNAs for a given gene. Combined with the optimized miR-E backbone, >90% of high-scoring SplashRNA predictions trigger >85% protein knockdown when expressed from a single genomic integration. SplashRNA can significantly improve the accuracy of loss-of-function genetics studies and facilitates the generation of compact shRNA libraries. The open source SplashRNA platform completes the RNAi toolkit to harness microRNA-based shRNAs for robust single-gene and multiplexed inducible and reversible target inhibition. Citation Format: Raphael Pelossof, Lauren Fairchild, Christina S. Leslie, Christof Fellmann. SplashRNA, a sequential classification algorithm for ultra-potent RNAi [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr LB-271. doi:10.1158/1538-7445.AM2017-LB-271

Evaluation of Face Detection Algorithms for the Bank Client Identity Verification

Abstract Results of investigation of face detection algorithms efficiency in the banking client visual verification system are presented. The video recordings were made in real conditions met in three bank operating outlets employing a miniature industrial USB camera. The aim of the experiments was to check the practical usability of the face detection method in the biometric bank client verification system. The main assumption was to provide a simplified as much as possible user interaction with the application. Applied algorithms for face detection are described and achieved results of face detection in the real bank environment conditions are presented. Practical limitations of the application based on encountered problems are discussed.

ENHANCEMENT OF FAST FACE DETECTION ALGORITHM BASED ON A CASCADE OF DECISION TREES

Abstract. Face detection algorithm based on a cascade of ensembles of decision trees (CEDT) is presented. The new approach allows detecting faces other than the front position through the use of multiple classifiers. Each classifier is trained for a specific range of angles of the rotation head. The results showed a high rate of productivity for CEDT on images with standard size. The algorithm increases the area under the ROC-curve of 13% compared to a standard Viola-Jones face detection algorithm. Final realization of given algorithm consist of 5 different cascades for frontal/non-frontal faces. One more thing which we take from the simulation results is a low computational complexity of CEDT algorithm in comparison with standard Viola-Jones approach. This could prove important in the embedded system and mobile device industries because it can reduce the cost of hardware and make battery life longer.