International Journal of Advanced Engineering Research and Science (IJAERS) [Vol-4, Issue-1, Jan- 2017] https://dx.doi.org/10.22161/ijaers.4.1.43 ISSN: 2349-6495(P) | 2456-1908(O) www.ijaers.com Page | 265 VLSI Implementation of Eye Detection System Dr. L. Padma Sree, G. Vijaya Bharghavi VNR Vignana Jyothi Institute of Engineering and Technology, Hyderabad, Telangana, India Abstract— The eye detection is one of the most challenging problems in many applications such as image processing, pattern recognition and computer vision. This paper introduces efficient vlsi implementation of eye detection system. Face detection is a very important part of the developed eye detection algorithm. Face detection is done by using skin detection method. Skin detection is an extraction of skin color pixels and regions from an image. In this method, first input image is converted from RGB image to YUV image. With YUV domain, the skin pixels in an image are extracted. Then Morphological operation is done by using erosion method through which noise of an image is also eliminated. From this image, face region is extracted by identifying the skin pixels. Then eye detection is done according to rules of human face proportion. The eye detection system is implemented using Verilog-HDL and simulation is done by using Xilinx ISE. Keywords— Face Detection, Eye Detection, RGB, YUV. I. INTRODUCTION Eye detection is one of the most challenging problems in disciplines such as image processing, pattern recognition and computer vision [1]. One of the major application for eye detection is iris detection. Other application of eye detection algorithm can be used in cell phone, security system, safe driving etc. Eye portion in an image is identified as low illumination, higher density edges, more contrast in comparison with rest part of the face. Face detection is a very important part of the developed algorithm. Face detection is done by using skin detection method [2]. Skin detection method is an extraction of skin color pixels and regions from an image. Different approaches that are used to detect human faces are feature based, appearance based, and color based. The feature based method detects a human’s face depends on human facial features [3]. Because of its complexity, this method requires lots of computing and memory resources. Color based method is more reasonable for applications that require low computational effort. In general, each method has its own advantages and disadvantages. More complex algorithm typically gives very high accuracy rate but also requires lots of computing resources [4]. Skin detection method is an effective method to detect face regions due to its low computational requirements and ease of implementation. Compared to the featured based method, the skin detection method required very little process [5]. This paper focuses on algorithm of the Eye Detection System by using Hardware Description Language. Software used in this implementation is Xilinx Synthesis Tool version 14.1. XST comes with provision of implementing the design using hardware description language. Verilog HDL has been chosen as implementing language as it has some additional features in designing as well as verification. Isim simulator is used for functional simulation of the design. Functional simulation is done using test bench for each and every block individually. In the test benches, images are required to be given as an input. Input images are stored in a text file which contains all pixel values. Matlab 2014a is used to convert an image into text file so that the pixel values can be passed in the design. Verilog HDL does contain system task that can read a text file. Read text file can be loaded into a memory defined inside the design. Simulation result is observed in the waveform available in the Isim simulator. Test bench is also written using Verilog HDL. II. PROPOSED ALGORITHM The input image is converted into text values using MATLAB program. Those values are stored in files and these are passed through test bench. Fig. 1. Eye Detection System