Fast Face Research Articles

A Fast Face Changing Method Adaptable to Various Postures

Image face-changing technology is influenced by skin color and posture. After changing faces, the contour information or skin color information of the user's face will be lost, resulting in low similarity between the changed image and the user's face in the source image. This paper introduces a rapid face-changing method that can adapt to various poses. Regardless of whether the facial poses in the source image and template image differ significantly, this method can retain the facial contours and skin tone in the user's source image without deformation, resulting in a high similarity between the transformed image and the face in the source image. Additionally, it offers low computational complexity, quick processing speed, and a small amount of computational effort.

Joint super-resolution-based fast face image coding for human and machine vision

Joint super-resolution-based fast face image coding for human and machine vision

Fast face swapping with high-fidelity lightweight generator assisted by online knowledge distillation

Fast face swapping with high-fidelity lightweight generator assisted by online knowledge distillation

FFSDF: An improved fast face shadow detection framework based on channel spatial attention enhancement

Recent shadow detectors excel on simple datasets but encounter difficulties with facial shadow images under complex lighting due to the lack of annotated shadow masks, varying shadow sizes, and imbalances between the target and background. This leads to training difficulties, reduced accuracy, and slower processing, posing a significant challenge for precise and fast detection framework development. We collected images and created Extended Yale Shadow Detection Dataset (EYSDD). In comparison to other datasets, this dataset includes additional manually annotated shadow masks, making it suitable for training convolutional neural networks. To address this problem, we propose incorporating Channel Spatial Direction-aware Spatial Context (CSDSC) module into Fast Shadow Detection Network (FSDNet). Additionally, we introduce Selective Attention Inverted Residual Bottleneck (SAIRB) with Selective Attention Mechanism (SAM). Furthermore, we integrate Detail Enhancement Module (DEM), which refines low-level features, into Fast Face Shadow Detection Framework (FFSDF). Finally, compared to other methods, our model surpasses the baseline method FSDNet and the advanced method EVP by 3.5% and 1.9% in terms of IoU, and 1.8% and 4.3% in terms of Dice score, respectively. Our model has only 4.31 M parameters and achieves a computing speed of 0.022 sec/image, demonstrating superior efficiency compared to other methods.

A fast face recognition system based on annealing algorithm to optimize operator parameters

ABSTRACT In order to improve the recognition rate and operational efficiency of the system, a method in which the image features compensation coefficients are optimized by using an improved simulated annealing algorithm is proposed. Firstly, eight computational factors with low computational complexity are given, which can be used to compensate image features. Secondly, the design flow of face recognition algorithm is presented. Thirdly, an improved simulated annealing algorithm is designed to solve the optimal combination of feature compensation coefficients in the face recognition system. Fourthly, the results of the feature compensation coefficients recommended by the improved simulated annealing algorithm are applied to the Efficient Face Recognition Algorithm (EFRA) in this paper, and verified on the simulation platform. Experiments show that the recognition rate can reach 100% when the training images are 6 in ORL. The proposed algorithm also performs well in MU_PIE dataset.

Rapid Face Asset Acquisition with Recurrent Feature Alignment

We present Re current F eature A lignment (ReFA), an end-to-end neural network for the very rapid creation of production-grade face assets from multi-view images. ReFA is on par with the industrial pipelines in quality for producing accurate, complete, registered, and textured assets directly applicable to physically-based rendering, but produces the asset end-to-end, fully automatically at a significantly faster speed at 4.5 FPS, which is unprecedented among neural-based techniques. Our method represents face geometry as a position map in the UV space. The network first extracts per-pixel features in both the multi-view image space and the UV space. A recurrent module then iteratively optimizes the geometry by projecting the image-space features to the UV space and comparing them with a reference UV-space feature. The optimized geometry then provides pixel-aligned signals for the inference of high-resolution textures. Experiments have validated that ReFA achieves a median error of 0.603 mm in geometry reconstruction, is robust to extreme pose and expression, and excels in sparse-view settings. We believe that the progress achieved by our network enables lightweight, fast face assets acquisition that significantly boosts the downstream applications, such as avatar creation and facial performance capture. It will also enable massive database capturing for deep learning purposes.

Research on Fast Face Recognition Method Based on Double Decision Subspace

Because a large number of labeled face data samples in special scenes need a large number of training samples with identity markers, and it is impossible to accurately extract the characteristics of small samples, a fast face recognition method based on double decision subspace is proposed. A feature recognition structure based on double decision subspace is constructed to preprocess the face image and separate the local features of several corresponding face images. The local binary pattern is used to extract the local texture features of the face, and the deep convolution network face fast recognition model is constructed. The convolution network is used to share the weight, pool, and downsampling to reduce the complexity of the model. The constructed recognition model is used to recognize the features of the face image, and the fast face recognition is effectively completed. The experimental results show that the designed method has high recognition accuracy, less average recognition time, and good recognition performance.

Effect of handgrip exercise on QTc interval during different phases of menstrual cycle

Menopause is a major change in a women’s life, it refers to a permanent cessation of menstruation and fertility. The fast face of modern life has induced many problems like stress, anxiety, insomnia etc. associated with serious health problems. The most common disorders associated with menstruation among younger women are dysmenorrhea, pre-menstrual syndrome, menorrhagia, metrorrhagia. Aims and Objectives: To record the effect of handgrip exercise on QTc and RR interval in different phases of menstrual cycle and to compare with the baseline ecg values. The study was conducted on 40 medical (I MBBS) students from Mysore Medical College and Research Institute, Mysore. In this study a total of 40 female students aged between 17-19 years from 1st MBBS studying at Mysore Medical Collage and Research Institute, Mysore were selected as per the inclusion and exclusion criteria. Informed consent was obtained from each student. Data about the duration of previous two menstrual cycles and the days of menstrual bleeding were noted from the subjects statement. Subject monitored basal body temperature was used to identify phases of menstrual cycle. Ovulation was indicated by a sustained increase in the basal body temperature of at least 0.3°C after LH surge.

Effect of handgrip exercise on QTc interval during different phases of menstrual cycle

Introduction: Menopause is a major change in a women’s life, it refers to a permanent cessation of menstruation and fertility. The fast face of modern life has induced many problems like stress, anxiety, insomnia etc. associated with serious health problems. The most common disorders associated with menstruation among younger women are dysmenorrhea, pre-menstrual syndrome, menorrhagia, metrorrhagia. Aims and Objectives: To record the effect of handgrip exercise on QTc and RR interval in different phases of menstrual cycle and to compare with the baseline ecg values. Materials and Methods: The study was conducted on 40 medical (I MBBS) students from Mysore Medical College and Research Institute, Mysore. In this study a total of 40 female students aged between 17-19 years from 1st MBBS studying at Mysore Medical Collage and Research Institute, Mysore were selected as per the inclusion and exclusion criteria. Informed consent was obtained from each student. Data about the duration of previous two menstrual cycles and the days of menstrual bleeding were noted from the subjects statement. Subject monitored basal body temperature was used to identify phases of menstrual cycle. Ovulation was indicated by a sustained increase in the basal body temperature of at least 0.3°C after LH surge.

Fast face detection using a unified architecture for unconstrained and infrared face images

Human face detection along with its localization is a difficult task when the face is presented in the cluttered scene in an unconstrained scenario that might be with arbitrary pose variations, occlusions, random backgrounds and infrared (IR) environment. This paper proposes a novel face detection method which can address some of these issues and challenges quite successfully during face detection in unconstrained as well as infrared environments. It makes use of Fast Successive Mean Quantization Transform (FastSMQT) features for image enhancement and feature representation to deal with illumination and sensor insensitiveness. A split up Sparse Network of Winnows (SNoW) with Winnow updating rule is then exploited to speed up the original SNoW classifier. Finally, the features and classifiers are combined together with skin detection algorithm for face detection in crowd image and head orientation correction of near infrared faces. The proposed face detector is robust in handling pose, occlusion, illumination, blur and low image resolution. The experiment is performed on six challenging and publicly available databases, viz. BIOID, LFW, FDDB, UFI, WIDER FACE and IIT Delhi near infrared. The experimental results depict that the proposed method outperforms traditional as well as some advanced methods in detecting unconstrained and infrared faces under challenging situations.

A fast face recognition based on image gradient compensation for feature description

To improve the efficiency of traditional face recognition techniques, this paper proposes a novel face recognition algorithm called Image Gradient Feature Compensation (IGFC). Based on the gradients along four directions in an image, a fusion algorithm and a compensation method are implemented to obtain features of the original image. In this study, gradient magnitude maps of a face image are calculated along four directions. Fusion gradients and differential fusion gradients are produced by fusing the four gradient magnitude maps of a face image in multiple ways, and they are used as compensation variables to compensate the appropriate coefficients on the original image and build IGFC feature maps of the original face image. Subsequently, IGFC feature maps are divided into several blocks to calculate the concatenated histogram over all blocks, which is in turn utilized as the feature descriptor for face recognition. Principal component analysis (PCA) is used to cut down the number of dimensions in high-dimensional features, which are recognized by the Support Vector Machine (SVM) classifier. Finally, the proposed IGFC method is superior to traditional methods as suggested by verification studies on YALE, ORL, CMU_PIE, and FERET face databases. When the LibSVM parameter was set to ‘-s 0 -t 2 -c 16 -g 0.0009765625’, the algorithm achieved 100% recognition on Yale and ORL data sets, 92.16% on CMU_PIE data sets, and 74.3% on FERET data sets. The average time for simultaneous completion of the data sets examined was 1.93 s, and the algorithm demonstrated a 70.71% higher running efficiency than the CLBP algorithm. Therefore, the proposed algorithm is highly efficient in feature recognition with excellent computational efficiency.

FAFD: Fast and Accurate Face Detector

Deep Neural Networks (DNN) have contributed a significant performance improvement in face detection. However, since most models focus only on the improvement of detection accuracy with computationally expensive structures, it is still far from real-time applications with a fast face detector. The goal of this paper is to improve face detection performance from the speed-focusing point of view. To this end, we propose a novel Fast and Accurate Face Detector (FAFD) to achieve high performance on both speed and accuracy performance. Specifically, based on the YOLOv5 model, we add one prediction head to increase the detection performance, especially for small faces. In addition, to increase the detection performance of multi-scale faces, we propose to add a novel Multi-Scale Image Fusion (MSIF) layer to the backbone network. We also propose an improved Copy-Paste to augment the training images with face objects in various scales. Experimental results on the WiderFace dataset show that the proposed FAFD achieves the best performance among the existing methods in a Speed-Focusing group. On three sub-datasets of WiderFace (i.e., Easy, Medium, and Hard sub-datasets), our FAFD yields average precisions (AP) of 95.0%, 93.5%, and 87.0%, respectively. Also, the speed performance of the FAFD is fast enough to be included in the group of speed-focusing methods.

The model of fast face recognition against age interference in deep learning

The model of fast face recognition against age interference in deep learning

The model of fast face recognition against age interference in deep learning

In order to overcome the low recognition efficiency of traditional anti-age face recognition methods, the paper proposes a new anti-age-disturbing face recognition modelling method based on deep learning. Firstly, build a standard face recognition information database and use this as the matching standard for face recognition. Then, construct a deep learning convolutional neural network, install the propagation process and training strategy of deep learning. Lastly, build an age discrimination model, a loss function and an objective function and solve it. On this basis, the facial features are extracted, after matching with the data in the standard database and similarity calculation, the final rapid facial recognition result is obtained. Experimental results show that the highest recognition accuracy of the designed face recognition model is 99.2%, and the recognition speed of the designed model is faster.

Abstract 13202: The Accuracy Of EMS Dispatch For Stroke Following The Introduction Of The Stroke Diagnostic Tool

Introduction: Early identification of stroke in the emergency call is key to timely reperfusion therapy. The Medical Priority Dispatch System (MPDS) is widely used and recent versions include a Stroke Diagnostic Tool (SDxT) using the FAST (Face, Arm, Speech) assessments. We conducted a retrospective observational study to examine whether introduction of the SDxT improved the identification of stroke/TIA in the emergency call in Victoria, Australia. Methods: We included all stroke dispatches (MPDS Code 28) for adults (aged ≥20 years) between 2008 and 2015. We linked this data to diagnoses from Ambulance Victoria, and the Victorian Department of Health and Human Services’ databases: the Victorian Emergency Minimum Dataset and the Victorian Admitted Episode Database. We examined the use of SDxT codes, and compared the proportion of stroke/TIA diagnoses before and after the introduction of the SDxT in September 2013. Results: Over the study period there were 108,917 Code 28 cases (average 13,600/year), 80% (n=86,906) were transported to an ED and 55% (n=60,334) were admitted to hospital. There were 3,141 diagnoses among the Code 28 cases, with 30% diagnosed with stroke/TIA. No changes was seen in the proportion of stroke/TIAs after the introduction of the SDxT (28.5% vs. 28.2%, p=0.23). The SDxT was used and complete in 63% (n=20,070) of calls. Patients with evidence of stroke on the SDxT were more likely to have stroke/TIA (partial evidence=29%, strong evidence=38% and clear evidence=40%) compared to those with no evidence (20%) and those with no/incomplete SDT (25%). Conclusions: Our results suggest a large proportion of stroke dispatches are not stroke/TIA, but evidence of stroke on the SDxT does identify marginally higher proportions. Further research is required to understand use of the tool, and public education on stroke assessments may be needed to realise its full potential.

3D Face Recognition Based on an Attention Mechanism and Sparse Loss Function

Face recognition is one of the essential applications in computer vision, while current face recognition technology is mainly based on 2D images without depth information, which are easily affected by illumination and facial expressions. This paper presents a fast face recognition algorithm combining 3D point cloud face data with deep learning, focusing on key part of face for recognition with an attention mechanism, and reducing the coding space by the sparse loss function. First, an attention mechanism-based convolutional neural network was constructed to extract facial features to avoid expressions and illumination interference. Second, a Siamese network was trained with a sparse loss function to minimize the face coding space and enhance the separability of the face features. With the FRGC face dataset, the experimental results show that the proposed method could achieve the recognition accuracy of 95.33%.

The Role of Emotional Content and Perceptual Saliency During the Programming of Saccades Toward Faces

Previous studies have shown that the human visual system can detect a face and elicit a saccadic eye movement toward it very efficiently compared to other categories of visual stimuli. In the first experiment, we tested the influence of facial expressions on fast face detection using a saccadic choice task. Face-vehicle pairs were simultaneously presented and participants were asked to saccade toward the target (the face or the vehicle). We observed that saccades toward faces were initiated faster, and more often in the correct direction, than saccades toward vehicles, regardless of the facial expressions (happy, fearful, or neutral). We also observed that saccade endpoints on face images were lower when the face was happy and higher when it was neutral. In the second experiment, we explicitly tested the detection of facial expressions. We used a saccadic choice task with emotional-neutral pairs of faces and participants were asked to saccade toward the emotional (happy or fearful) or the neutral face. Participants were faster when they were asked to saccade toward the emotional face. They also made fewer errors, especially when the emotional face was happy. Using computational modeling, we showed that this happy face advantage can, at least partly, be explained by perceptual factors. Also, saccade endpoints were lower when the target was happy than when it was fearful. Overall, we suggest that there is no automatic prioritization of emotional faces, at least for saccades with short latencies, but that salient local face features can automatically attract attention.

STOCK MARKET LINKAGES AND CAUSAL RELATIONSHIPS: EMPIRICAL INVESTIGATION OF EM7 ECONOMIES

Diffusion of information in the present era has become very fast, whether it is related to natural phenomena or human activities. Due to the technological advancement and fast face globalisation and liberalisation, events happening in financial markets are no exception, especially due to electronic stock exchanges and free flow of capital and financial information across borders. The present study aims to examine return patterns and find inter linkages/integration among the stock markets of seven largest emerging economies popularly known as EM7 (India, China, Russia, Brazil, Indonesia, Mexico and Turkey) by examining the monthly return data from Jan 2010 to Dec 2019. The study used descriptive analysis, correlation analysis, regression analysis and causality test to attain its objectives. The results indicate that EM7 stock markets are not interlinked, suggesting markets are quite segmented and there is scope for fund managers and both international and domestic investors to reap the advantages of portfolio diversification and mitigate the risks associated with their investments.

Abstract P583: Comparison of FAST-S to BEFAST-S for Pediatric Stroke: A Comprehensive Stroke Center Retrospective Analysis

Objective: To assess the demographic characteristics and presenting symptoms in a cohort of pediatric patients with ischemic stroke and develop a tool to improve recognition of ischemic stroke at the institution and in the community. Methods: We conducted a retrospective review of children aged 1 month to 18 years with radiologically confirmed ischemic stroke admitted to the University of Kentucky between January 2013 and June 2020. Characteristics such as age, sex, and diagnosis were recorded. The FAST (Face, Arms, Speech) and BEFAST (Balance, Eyes, Face, Arms, Speech) mnemonics were applied to all patients. We then applied the FAST-S tool and used the novel BEFAST-S. FAST-S adds new onset seizure to the previously validated tool. Results: For the study period, we identified 158 cases. Most patients were ≥ 3 years of age (n=113) and most were male (n=96). About one third (31%, n=50) of presentations concerning for stroke had diagnosed ischemic stroke at time of discharge; the majority were ≥ 3 years of age (n=31). Strokes were commonly associated with trauma (24%, n=12) and infection (22%, n=11). Common mimics for stroke included seizure in the absence of stroke (30%, n=33), migraine (18%, n=20), and conversion disorder (13%, n=14). FAST and BEFAST each identified the same 30 patients as having ischemic stroke (60%), with 26 patients being ≥ 3 years of age. Inclusion of new onset seizures (FAST-S/BEFAST-S) correctly identified one additional patient ≥ 3 years of age (87%, n=27) and nine additional patients under 3 (n=19, 78% versus n=4, 21%). New onset seizure reduced the number of missed strokes regardless of age from 40% (n=20) to 16% (n=8). Conclusions: Although BEFAST improves the sensitivity of identification of ischemic strokes in adults, it did not identify additional strokes beyond FAST in our pediatric population. Inclusion of new onset seizure in addition to either FAST or BEFAST, however, reduced the number of missed strokes by at least half, particularly in children < 3 years of age. Thus, education of FAST-S or BEFAST-S may be useful for providers and emergency services.

DC-EDN: densely connected encoder-decoder network with reinforced depthwise convolution for face alignment

High accuracy and fast face alignment algorithms play an important role in many face-related applications. Generally, the model speed is inversely related to the number of parameters. We construct our network based on densely connected encoder-decoders, which is an efficient method to balance the parameter number and localization results. In each encoder-decoder, we introduce stacking depthwise convolution and depthwise feature fusion within the same channel, which greatly improves the performance of depthwise convolution and reduces the number of model parameters. In addition, we enhance the mean square loss function by assigning different penalty weights to each coordinate according to the distance to the position corresponding to the maximum value in the label heatmap. Experiments show that the model with the improved loss function obtains better localization results. In the experiment, we compare our method to state-of-the-art methods based on 300W and WFLW. The localization error is 2.76% with the common subset of 300W and the model size (0.7M) is small and even utilizes approximately 1% of the number of parameters of the other models. The dataset and model based on WFLW are publicly available at https://github.com/iam-zhanghongliang/DC-EDN .