An FPGA-based Parallel Hardware Architecture for Real-time Eye Detection

Abstract

Abstract-Eye detection is widely used in applications, such as face recognition, driver behavior analysis, and human-computer interaction. However, it is difficult to achieve real-time performance with software-based eye detection in an embedded environment. In this paper, we propose a parallel hardware architecture for real-time eye detection. We use the AdaBoost algorithm with modified census transform(MCT) to detect eyes on a face image. We parallelize part of the algorithm to speed up processing. Several downscaled pyramid images of the eye candidate region are generated in parallel using the input face image. We can detect the left and the right eye simultaneously using these downscaled images. The sequential data processing bottleneck caused by repetitive operation is removed by employing a pipelined parallel architecture. The proposed architecture is designed using Verilog HDL and implemented on a Virtex-5 FPGA for prototyping and evaluation. The proposed system can detect eyes within 0.15 ms in a VGA image.

Pesquisa de Posição Ocular por Correpondência de Modelos para Renderização Baseada em Fóvea com FPGA

“…Em relação aos recursos usados, o compilador Quartus II reportou um uso de 2.292 unidades lógicas, o que corresponde a 7% do total disponível, se mostrando um design muito compacto. Para um outro trabalho disponível na literatura, Kim et al (2012) mostram um uso de aproximadamente 39.914 unidades lógicas, que para a FPGA escolhida por eles corresponde a um uso de 76% das unidades lógicas disponíveis.…”

Eye Gaze Techniques for Human Computer Interaction: A Research Survey

“…In this method the eye position is detected via a recursive process. This method can detect the eye accurately, but it is computationally expensive, demanding and high contrast image is required [21][22][23]. In principle, the template matching technique is weak in pattern rotations and scale changes.…”

“…The appearance based methods [15, 27-29, 33, 40] may include pupil and purkinje or pupil and glint vector for eye detection. Shape based methods [9,20,27,32,35] Template based methods [21,22] use deformable template or between the eyes methods [8]. The other methods can use combination of the two or some of the above mentioned methods.…”

“…One IR [24,27] Interactive Trackers [19,21,27,30,32,34] DASHER use EyeTech"s Quick Glance eyetracker for eye typing [12]. Free head gaze tracking system with the eye positioning unit mounted on the display and the gaze tracking unit placed to detect the user"s gaze position [30].…”