FPGA Prototypes

Demonstration of Real-Time Computer Vision Algorithms on FPGA platform [More Info]

○ Implemented a prototype for a custom accelerator for Computer Vision tasks (filters, face detection, etc.)

○ Optimized hardware architecture for parallel processing. Implemented using Verilog HDL and C

○ Platform: Spartan 6 LX150t Industrial Video Processing Kit

○ Hardware Description Language: Verilog using Xilinx design suite (ISE and EDK)

○ Embedded Processor: Microblaze

○ Programming Language: C

Real-Time Coin Recognition Demo on a Spartan 6 Industrial Video Processing FPGA Board 📰

○ Implemented a prototype for a custom accelerator Coin Recognition between euro and counterfeit coins.

○ Design of a enclosing system with LED lights for lighting, and camera for visual processing.

○ Optimized hardware architecture based on Support Vector Machines and Edge Images.

○ Prototype is capable of processing videos of coins in real-rime (at the camera capturing rate)

○ A more general coin recognition system can be build on top of these system by combining binary SVM classifiers in a tree structure

Real-Time Face Detection Demo on a ML505 Virtex 5 FPGA Board

○ This project focuses on the design of a hardware system for the purpose of performing upright frontal view face detection on an image frame.

○ The algorithm uses a Support Vector Machine

○ The hardware accelerator processes the data in parallel and outputs the bounding boxed on the detected objects

○ Platform: Virtex 5 LX110t ML505 platform

○ Hardware Description Language: Verilog using Xilinx design suite (ISE and EDK)

○ Embedded Processor: Microblaze

○ Programming Language: C

Talk on our proposed Support Vector Machine Hardware Architecture [Kyrkou 2012, IEEE TC]

Christos Kyrkou, and Theocharis Theocharides, "A Parallel Hardware Architecture for Real-Time Object Detection with Support Vector Machines", IEEE Transactions on Computers, vol.61, no.6, pp.831-842, June 2012.