VNU-UET Repository

Accurate and Low Complex Cell Histogram Generation by Bypass the Gradient of Pixel Computation

Ho, Huy Hung and Nguyen, Ngoc Sinh and Bui, Duy Hieu and Tran, Xuan Tu (2017) Accurate and Low Complex Cell Histogram Generation by Bypass the Gradient of Pixel Computation. In: The 4th NAFOSTED Conference on Information and Computer Science (NICS), 24-25 November 2017, Hanoi, Vietnam.

[img] PDF
Download (1MB)


Histogram of Oriented Gradient (HOG) is a popular feature description for the purpose of object detection. However, HOG algorithm requires a performance system because of its complex operation set. Especially, the cell histogram generation is one of the most complex part, it uses inverse tangent, square, square root, floating point multiplication. In this paper, we propose an accurate and low complex cell histogram generation by bypass the gradient of pixel computation. It employs the bin’s boundary angle method to determine the two quantized angles. However, instead of choosing an approximate value of tan, the nearest greater and the nearest smaller of each tan values from ratio between pixel’s derivative in y and x direction are used. Magnitude of two bins are solutions of a system of two equations, which represent the equality of the gradient of a pixel and its two bins in both vertical and horizontal direction. The proposed method spends only 30 addition and 40 shift operations to identify two bins of a pixel. Simulation results show that the percentage error when reconstructing the difference in x and y direction are always less than 2% with 8-bit length of the fractional part. Additionally, manipulating the precision of gradient magnitude is very simple by pre-defined sine and cosine values of quantized angles. Synthesizing the hardware implementation presents that its area cost is 3.57 KGates with 45nm NanGate standard cell library. The hardware module runs at the maximum frequency of 400 MHz, and the throughput is 0.4 (pixel/ns) for a single module. It is able to support about 48 fps with 4K UHD resolution.

Item Type: Conference or Workshop Item (Paper)
Subjects: Electronics and Communications
Electronics and Communications > Electronics and Computer Engineering
Information Technology (IT)
Divisions: Faculty of Electronics and Telecommunications (FET)
Key Laboratory for Smart Integrated Systems (SISLAB)
Depositing User: Prof. Xuan-Tu Tran
Date Deposited: 08 Dec 2017 07:21
Last Modified: 08 Dec 2017 07:21

Actions (login required)

View Item View Item