Hardware implementation of template matching algorithm and its performance evaluation
B. Sowmya Satish, P. Jayakrishnan · 2017 International conference on Microelectronic Devices, Circuits and Systems (ICMDCS) · 2017
Object Recognition is used to identify the object in the form of a video sequence or an image inside a video sequence or an image. The main objective of the Object Recognition is to correctly find the objects in the image or video sequence. There are different techniques are available to match the area. The goal is to identify the similar object or exact object in the input image. Template Matching Algorithm is used to detect and recognize an object using Optimized Sum of Absolute Differences (OSAD). In the proposed method, it does not suffer from the illumination issues. It got high performance compare to other technique. It is verified in Matlab, Simulink and Verilog. For the hardware simulation, Copies the output image pixel values from Verilog to text file and give it to the matlab and displays the detected object on the image. It is verified by using the FPGA Board (Arria 10 Device Family) and got the output less than 4ms.