A Highly Accurate and Parallel Vision MLP FPGA Accelerator based on FP7/8 SIMD Operations.
Mari YasunagaJunnosuke SuzukiMasato WatanabeKazushi KawamuraThiem Van ChuJaehoon YuMasato MotomuraPublished in: MCSoC (2023)
Keyphrases
- highly accurate
- parallel implementation
- field programmable gate array
- processing elements
- massively parallel
- parallel architectures
- hardware implementation
- parallel computing
- parallel processing
- hardware architecture
- processor array
- parallel architecture
- single instruction multiple data
- parallel hardware
- real time
- multilayer perceptron
- high quality
- message passing interface
- image processing algorithms
- embedded systems
- processing units
- neural network
- high speed
- computer vision
- high accuracy
- capable of producing
- distributed memory
- artificial neural networks
- parallel computers
- multi layer perceptron
- vision system
- compute intensive
- highly parallel
- accurate models
- fpga implementation
- neural classifier
- shared memory
- real time image processing
- smart camera
- array processor
- mesh connected
- parallel computation
- signal processing
- efficient implementation
- int j imaging syst technol
- high end
- parallel programming
- parallel machines
- computer architecture