This document presents a multi-core processor architecture on an FPGA for image feature extraction and classification. It shows that an FPGA-based multi-core processor can process images faster than a single-core processor, with correlation results that are comparable to MATLAB. The FPGA implementation was able to achieve image processing times that were 3.419 times faster than a single-core processor. This FPGA-based multi-core architecture allows for flexible and high-performance image processing.
A Parallel, Energy Efficient Hardware Architecture for the merAligner on FPGA...
IEEEFYP2014poster Track 5 Vincent Yeong Chun Kiat
1. An FPGA-Based Multi-core Processor Architecture for
Image Feature Extraction and Classification
Name: Vincent Yeong Chun Kiat
IC Number: 900528-14-5683,UPM matrix number: 156434,Email: vincentkarl90@gmail.com
2014 IEEE Malaysia Final Year Project Competition
Insert
your
picture
here
1
2
3
1
2
3
Display captured image
Correlation result from CPUs
Time taken to process
ABSTRACT
Resize to
256x128
pixels
Apply GLCM
and Correlation
Apply GLCM
and Correlation
Finalized
Correlation and
Classification
Capture
KEY[0] Read out
SDRAM
Timer
Output
CPU 1
( 128x128
pixels)
CPU timer
High Level View of System Architecture:
A. Comparison on Correlation between FPGA
and MATLAB on flooring types: Average correlation
FPGA 0.615813
MATLAB 0.751162
Difference 0.135349
Percentage Difference(%)
= (0.135349
0.751162)×100%
= 18.0286%
B. Performance Evaluation:
FPGA
processors
Time Taken for
Execution (ms)
Single-core 8861.667
Multi-core 2591.633
RESULTS & DISCUSSION
Rate(Execution)
= (8861.667 𝑚𝑠
2591.633 𝑚𝑠)
= 3.41934 Times Faster
Execute different image processing algorithms on soft-core processor
FPGA-based multi-core processor is a better option: » fast operation and comparable results
» easily adaptable for other IP applications
CONCLUSION
0
0.2
0.4
0.6
0.8
1
1.2
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58
Correlation
Number of Sample Points
Correlation on Wood
Matlab_correlation Real time Correlation
Image processing computationally intensive operation
requires immense resources in CPU and memory throughput.
high parallelism in image processing suitable FPGAs.
FPGA-based real time image processing using mixed HW/SW co-design on multi-core processor
platform is introduced:
• Exploit parallelism in image processing algorithms
• Utilize FPGA technology → use of multi-core processor to speed up operations
• Fast prototyping
Output:
CPU 2
( 128x128
pixels)
C. Flexible Adaptability
Functional Block
Diagram:
Concurrent design
METHODOLOGY