Font Size: a A A

The SOC Design And Verification Of Eye Gaze Tracking System

Posted on:2014-11-13Degree:MasterType:Thesis
Country:ChinaCandidate:Z W HuFull Text:PDF
GTID:2268330401459011Subject:Electronics and Communications Engineering
Abstract/Summary:PDF Full Text Request
SOC(System on Chip) integrate the microprocessor, digital module, analog moduleand memory into a single chip, facilitating the speed, power consumption and functionvariety. Therefore, SOC technique and methodology is becoming the mainstream of ICdesign. SOPC (System on Programmable Chip) is an efficient and flexible solution whichadopts on-chip bus and reuse methodology to integrate all hardware IP core and softwarefunction modules into a single FPGA (Field-Programmable Gate Array).This paper develops the design and implementation of Eye Gaze tracking SOC system,based on on-chip bus and IP reuse methodology. Analyzing the various plans and datapathof the initial algorithm, this paper focus on implementation and verification on SOPCplatform.The contents and major contributions of this paper are listed as follow:(1)Presents the fundamental concept and methodology of SOC design in the field of ICdesign. Analyzes the market and application perspective of Eye Gaze tracking System.(2)The basic principles and flow of Eye Gaze tracking System based on pupil cornealglint are introduced and the algorithm is divided into sub-modules, constituting thearchitecture of SOC system. With the fully analysis of Adaboost eye localization algorithm,the quick generator for integral image based on vector moving and hybrid cascade classifierare proposed.(3)Implements the ellipse fitting for eye pupil based on RANSAC and analyzes thedataflow of RANSAC algorithm, dividing into three sub-modules which are pseudo numbergenerator, quick matrix inversion and error accumulator.(4)At last, the integration and verification of Eye Gaze tracking on SOPC areintroduced. It’s demonstrated by the result of sample image that the SOC system highlycorresponds to the initial algorithm on PC, keeping the unvarnished feature of prototypealgorithm. The Eye Gaze tracking system on SOPC platform runs at11frames per second, achieving the real time performance basically. To conclusion, the work introduced in thispaper demonstrates the superiority of SOC technique and methodology.
Keywords/Search Tags:SOC, Gaze tracking, on-chip bus, IP reuse, FPGA
PDF Full Text Request
Related items