All news »

Your Name:

Your Email:



verification image, type it in the box
Put in the spam-code here:
+7 (495) 790 8117
Human Face Identification Using 3D Neural Representation PDF Print E-mail
We study neural network approach for stereo image face recognition. Human Face Identification Using 3D Neural Representation, 3D Ident Person
We study neural network approach for stereo image face recognition. We consider the accepted leader in the field of 3D face recognition is A4Vision company (
Their approach is in projecting structured light on the surface using infrared lattice (invisible to the human eye) , that is then taken by the videocamera in the infrared spectrum and according to the distortions of the lattice caused by the curvature of the face the model of the surface is built. After this is done the system measures different features of the face which are then considered to be a personal "code" and then introduced into the database and later used for matching.
The approach we study is different. Instead of structured light, we want to use common stereo vision for height map modelling as a continious function z = f(x,y) represented by a neural network. The particular face is thus represented in neural network coefficients. In this approach full face surface is stored in the database (not just face "code") in the compact form of neural network weights that allows more accurate surface mathing. In addition iterative neural network algorithm allows warps measurements, that is face comparison under different non-linear face distortions caused by mimics. The project is on the prototyping phase.
Our products
Sample Image PWNBinaryTrainer 1.0

PWNBinaryTrainer 1.0 is a simple GPGPU console application designed to bring alternative to training of the neural networks in expensive commercial packages. Training process takes now minutes instead of hours and the accuracy is similar or better ...


Sample Image PWNLIB 1.1

Software C++ library for neural network processing and training. GPGPU via NVIDIA CUDA and Intel-MKL performance boosting is available now.


Our services

Custom coding of math algorithms

Even though modern tools exist that greatly simplify math algorithms coding (e.g. MATLAB), there are cases when it make sense to outsource this work to us:

Data visualization

We offer 1D,2D,3D data visualization (real time if necessary) by utilizing the latest dimension reduction techniques, linear and non linear PCA-related technologies, volume visualization, custom ray-tracing and radiosity algorithms. We offer fast OpenGL or GPGPU / CUDA implementations.

Data/Image processing

There are common computer vision libraries such as Intel OpenCV that one can take as a starting point for their algorithms implementation.

Recognition and Clustering

Typical task for neural networks is recognition and clustering. Send us your task description and we will provide you with a proposal on time and budget for your particular system implementation.

Embedded programming of intellectual algorithms

Embedded intellectual algorithms must be very simple, fast, and reliable. Pawlin Technologies currently works on getting the maximum from the limited hardware capabilities in term of pattern recognition, computer vision.

Neural Network Programming

We reveal new capabilities of using data and information for the goals of your business. Now many tasks historically done by humans can be left to intellectual software and hardware by using some of the many existing AI technologies customized for your particial needs.
All services »
Help us to be more informative, please, classify yourself: