• Download GPUNN_demo - 584.61 KB (release build requiring CUDA and 120dpi)
• Download GPUNN_GUI - 510.12 KB (the GUI source code)
• Download GPUNN_kernel - 2.78 KB (the Neural Network core)

Introduction

An Artificial Neural Network is an information processing method that was inspired by the way biological nervous systems function, such as the brain, process information. It is composed of a large number of highly interconnected processing elements (neurons) working in unison to solve specific problems. Neural Networks have been widely used in "analogous" signal classifications, including handwriting, voice and image recognitions. Neural network can also be used in computer games. It enables games with the ability to adaptively learn from player behaviors. This technique has been used in racing games, such that opponent cars controlled by computer can learn how to drive from human players.

Since a Neural Network requires considerable number of vector and matrix operations to get results, it is very suitable to be implemented in a parallel programming model and run on Graphics Processing Units (GPUs). Our goal is to utilize and unleash the power of GPUs to boost the performance of a Neural Network solving handwriting recognition problems.

This project was originally our graphics architecture course project. We ran on GPU the same neural network described by Mike O'Neill in his brilliant article "Neural Network for Recognition of Handwritten Digits".

• Download source files - 100 Kb
• Download demo project - 167 Kb

Introduction

What is DXF?

Drawing Interchange Format (DXF) files enable the interchange of drawings between AutoCAD and other programs. DXF files can be either ASCII or binary formats. Because ASCII DXF files are more common than the binary format, CadLib uses ASCII DXF format.

What is CadLib?

The CadLib is not a Computer Aided Design (CAD) program. It is a tool for creating DXF files that are used in the CAD programs. It consists of two parts. One of them is a Dynamic Link Library to create the DXF file. The other part is the programming interface. It is a class that integrates the cadio.dll functions. It can be used in Microsoft Visual C++ projects. In addition, the cadio.dll can be used in other Win32 programs.

Why use CadLib?

In some programs, it is needed to create a drawing output for use in other programs such as AutoCad. For example, in a "Building Detail Sheet Generator Program", the program needs to create a drawing output. And the most standard format for communicating drawing data is DXF.

DXF file structure

The DXF format is a tagged data representation of all the information contained in a drawing file. Tagged data means that each data element in the file is preceded by an integer number that is called a group code. A group code's value indicates what type of data element follows. This value also indicates the meaning of a data element for a given object (or record) type. Virtually all user-specified information in a drawing file can be represented in DXF format. (from AutoCad's DXF reference)

A DXF file consists of some sections. Each section has some drawing data in itself. The CadLib uses the following sections:

1. HEADER
2. TABLES
3. BLOCKS
4. ENTITIES

The main reference for DXF file structure that is used for CadLib is the AutoCad's DXF reference. You can find more information about DXF file structure here.