Spread and Iterative Search: A High Quality Motion Estimation Algorithm for High Definition Videos and Its VLSI Design

Abstract

This paper presents the Spread and Iterative Search (S&IS) motion estimation algorithm, which uses a random spread evaluation together with a central iterative evaluation to avoid local minima falls and to increase the image quality for high definition videos. Considering Full HD videos, S&IS reached an average PSNR gain of 1.41dB when compared to Diamond Search (DS), with an increase of about four times in the number of evaluated blocks. When compared to Full Search (FS), the S&IS achieved an average PSNR loss of 1.56 dB, evaluating 73 times less blocks than FS. An efficient architecture for the S&IS algorithm is also presented in this paper.The architecture was designed targeting in real time processing (30 frames per seconds) for QFHD videos (3840×2160 pixels). The architecture was described in VHDL and synthesized for and Altera Stratix 4 FPGA and for ST90nm standard cells technology. Booth syntheses show that the architecture is able to process QFHD frames in real time. The standard cells version is able to reach also a good trade-off among area, memory and power consumption, processing QFHD videos with 62.2 mW.

ES&IS: Enhanced Spread and Iterative Search hardware-friendly motion estimation algorithm for the HEVC Standard

“…On our previous work [6] an architecture for the S&IS algorithm was presented. This architecture could achieve a processing rate of 149 HD 1080p frames per second using a block size of 16x16 and processing a block in 174 cycles.…”

“…This paper presented the ES&IS hardware-friendly ME algorithm, which is an upgrade of the algorithm presented in our previous work [6]. This algorithm was designed focusing on some characteristics that are important on hardware design such as non-need of a memory to store information of MV from previous encoded frames and a limited number of operations per encoded block.…”

“…For each evaluation a total of 60 frames of each video sequence had been encoded. For both S&IS and ES&IS the number of spread positions used in the evaluation was 96 (24 per quadrant) because in our previous work [6] it was demonstrated that this value of N maximize the algorithm efficiency.…”

“…The S&IS was compared to FS and Diamond Search (DS) and obtained a good tradeoff between complexity and video quality. Through the software evaluation presented in [6] it is not possible to analyze the efficiency of the S&IS algorithm inside a specific video coding standard.…”

“…In our previous work [6] the Spread and Iterative Search (S&IS) algorithm and its architectural design were presented. The S&IS insert randomness in the ME process to avoid local minima falls.…”