The human face exhibits a high level of complexity when it is regarded as a multidimensional visual model, leading to face recognition systems that require difficult and extensive computations for coding and decoding the face images. A well-established approach in this regard is based on using principle component analysis (PCA) for both feature extraction and face recognition, known as the eigenface approach. This technique, despite a good recognition rate, suffers from the disadvantage of high computation cost due to the complexity of the PCA algorithm. In this paper, we use a geometrical approximated PCA (gaPCA) algorithm for computing the eigenfaces for three different datasets. The face recognition task is performed using a similarity score based on the inverse Euclidean distance for the first two datasets and using a nerual network in the third case. All the results are compared to the case where standard PCA is used. These accuracy results show that gaPCA represents a viable alternative to the classical statistical approach for computing the principal components.

Face Recognition Using Eigenfaces, Geometrical PCA Approximation and Neural Networks

The assumption that the number of training samples is less than the number of pixels in a face image is essential for conventional eigenface-based face recognition But recently, it has become impractical for massive face image collections A parallel processing method using distributed eigenfaces is presented A massive face image set was divided into a bunch of small subsets that satisfied the assumption of conventional approaches Eigenfaces were extracted from the subsets and stored in a cloud system Face recognition was performed by parallel processing using the distributed eigenfaces in the cloud system A face recognition system was implemented in the Hadoop system Various experiments were performed to test the validity of the distributed eigenface-based approach The experimental results show that, compared to conventional methods, the implemented distributed face recognition system worked well for large datasets without significant performance degradation

Distributed eigenfaces for massive face image data

Large-scale floods are one of the major events that impact the national economy and people’s livelihood every year during the flood season. Predicting the factors of flood evolution is a worldwide problem. We use the two-dimensional Saint-Venant equations as an example and for high-performance computing in modelling the flood behavior. Discretization of the two-dimensional Saint-Venant equations with initial and boundary conditions with the finite difference method in the explicit leapfrog scheme is carried out. Afterwards, we employed a large-scale heterogeneous parallel solution on the “SunRising-1” supercomputer system using MPI, OpenMP, Pthread, and OpenCL runtime libraries. On this basis, we applied communication/calculation overlapping and the local memory acceleration to optimize the performance. Finally, various performance tests of the parallel scheme are carried out from different perspectives. We have found this method is efficient and recommend this approach be used in solving systems of partial differential equations similar to the Saint-Venant equations.

https://www.mdpi.com/2076-3417/12/11/5671/pdf?version=1654498941

Heterogeneous Parallel Implementation of Large-Scale Numerical Simulation of Saint-Venant Equations

The identification of people by means of biometric features has been an active research topic before computers started being used for such tasks. The scale of problems than can be addressed by the use of personal computers or computer clusters has grown but there are still real-world scenarios where the biometric features database cannot be processed in real time. This paper tackles the use of graphics processing units (GPUs) which are high performance and cost effective computing elements, as building blocks of large-scale identification systems. A review of the existing GPU based systems is introduced focusing on GPU based fingerprint matching systems. The impressive results obtained by GPU-based systems are only possible by applying a careful biometric matching algorithm redesign, task granularity, asynchronous memory transfers and task overlapping techniques.

/pdf/gpu-processing-for-biometric-big-data-based-identi-ication-47g7lq7uf2.pdf

GPU Processing for Biometric Big Data Based Identi.ication. Why and what for

Face recognition is a recent active research field that regroups many disciplines: image-processing, security, face detection, computer vision...etc. Due to their complexity, face recognition applications require large computations and memory space. The emergence of modern GPU with high parallel computational capabilities had permitted overcoming this issue, and exploiting the inherent of these applications. Although, several ongoing research works have made some progress on making the recognition process faster and more accurate in real time application, nevertheless, there are still more to do. In this paper, we will present a survey on face recognition accelerating methods, precisely on Graphic Processing Units.

GPU-based Face Recognition Acceleration Techniques: A Survey

Face recognition has many real world applications including surveillance and authentication. Due to complex and multidimensional structure of face it requires huge computations therefore fast face recognition is required. One of the most successful appearance based techniques for face recognition is Principal Component Analysis (PCA) which is generally known as eigenface approach. It suffers from the disadvantage of higher computation cost, despite its better recognition rate. With the increase in number of images in training database and also the resolution of images, the computational cost also increases. In this paper, we present a CUDA implementation of eigenface approach for face recognition. The proposed algorithm has shown a 5× speedup in training phase.

Parallel implementation of eigenface on CUDA

Heat conduction is natural phenomenon which is governed by three dimensional, transient partial differential equation. The partial differential equation can be solved by many numerical method such as finite difference method, finite volume method, finite element method etc. These methods require heavy computation to solve the system of algebraic equations. Graphics processing unit (GPU) can be used to handle the computation of CPU as a co-processor so the GPU will save a lot of time for computation. The dominant proprietary framework for GPU computing is CUDA, provided by NVidia. It can be used to solve computation intensive task. The objective of this work is to develop a heat conduction code on CUDA platform, which will solve the system of algebraic equations using GPU framework.

Manik R. Kawale

Papers

Parallel implementation of eigenface on CUDA

Development of 3D-CFD code for heat conduction process using CUDA