General-purpose 128 128 SIMD processor array with integrated image sensor

نویسنده

  • P. Dudek
چکیده

Introduction: In some computer vision applications, especially where high computational performance is required together with low power consumption, it is beneficial to use ‘vision chips’—devices that combine image sensing and processing on a single silicon die. Application-specific vision chips can efficiently perform some uncomplicated low-level image processing tasks, such as filtering, edge detection, or motion detection, often using analogue circuitry operating in pixel-parallel fashion. But to combine different operations to form a more complex algorithm, and to provide flexibility to use a chip for various algorithms, a software-programmable solution must be employed. The required flexibility can be provided by general-purpose SIMD processor arrays. We have been working towards combining the two paradigms: a general-purpose SIMD array and analogue processor-per-pixel vision systems [1–3]. Our first proof-of-concept ‘analogue SIMD vision chip’ [1] (SCAMP— switched current analogue matrix processor), was a 21 21 processor array. An improved, 39 48 chip version (SCAMP-2) was subsequently developed [2] followed recently by a 128 128 (SCAMP-3) device [3]. In this Letter we overview the SCAMP-3 chip design and present measured performance=accuracy data and experimental image processing results.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

SCAMP-3: A Vision Chip with SIMD Current-Mode Analogue Processor Array

In this chapter, the architecture, design and implementation of a vision chip with general-purpose programmable pixel-parallel cellular processor array, operating in single instruction multiple data (SIMD) mode is presented. The SIMD concurrent processor architecture is ideally suited to implementing low-level image processing algorithms. The datapath components (registers, I/O, arithmetic unit...

متن کامل

Implementation of SIMD vision chip with 128×128 array of analogue processing elements

This paper presents the latest implementation of the SIMD Current-mode Analogue Matrix Processor architecture. The SCAMP-3 vision chip has been fabricated in a 0.35μm CMOS technology and comprises a 128×128 general-purpose programmable processor-perpixel array. The architecture of the chip is overviewed and implementation issues are considered. The circuit design of the analogue register is pre...

متن کامل

A 39x48 GENERAL-PURPOSE FOCAL-PLANE PROCESSOR ARRAY INTEGRATED CIRCUIT

This paper presents the implementation of a generalpurpose programmable vision chip, with a 39×48 SIMD processor-per-pixel array, fabricated in a 0.35μm CMOS technology. The chip employs Analogue Processing Elements to achieve cell density of 410 cells/mm. The array operates at 1.25MHz with power consumption of 12μW/cell and executes low-level image-processing algorithms in real-time. Chip arch...

متن کامل

Vision Sensor with a SIMD Processor Array in a Vertically Stacked 3D Integrated Circuit Technology

We present a design of a vision sensor device, implemented in a three-dimensional (3D) silicon on insulator (SOI) 150nm CMOS technology. The proof-of-concept device contains an image sensor array on one layer, and a pitchmatched array of 32x32 pixel-parallel processors, distributed over two further layers. The processor array uses mixed-mode processing elements and operates in SIMD (Single Inst...

متن کامل

Parallelization of 3-D Range Image Segmentation on a SIMD Multiprocessor

The paper presents a parallelized algorithm for seg-mentation of 3-D range images. The segmented image is useful as an input to higher level image processing tasks. The algorithm is implemented on a mesh connected multiprocessor machine, the MasPar series of massively parallel computers. A 2-dimensional hierarchical data distribution scheme is used to allocate the image pixels on the machine. T...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2000