Abstract:With the development of Internet and cloud computing, multimedia data, such as images and videos, has become one of the most common data types being processed. As the scale of multimedia data being still increasing, it is vitally important to efficiently extract useful information from such a huge amount of multimedia data. However, due to the complexity of the core algorithms, multimedia retrieval applications are not only data intensive but also computationally intensive. Therefore, it has been a major challenge to accelerate the processing speed of such applications to satisfy the real-time requirement. As Graphic Processing Unit (GPU) has entered the general-propose computing domain (GPGPU), it has become one of the most popular accelerators for the applications with real-time requirements. In this paper, we parallelize a widely-used image retrieval algorithm called SURF on GPGPU, which is the core algorithm for many video and image retrieval applications. We first analyze the parallelism within SURF to guarantee that there are sufficient tasks being mapped to the large-scale computation resources in GPGPU. We then exploit some inherent GPGPU characteristics, such as 2D memory, to further boost the performance. Finally, we provide some optimization to the cooperation between CPU and GPGPU, which is generally ignored in previous designs. Experimental results show that our parallelization and optimization achieve a throughput of 340.5 frames/s on a NVIDIA GTX295 GPGPU, which is 15X faster than the maximal optimized CPU version. Compared to CUDA SURF, a state-of-the-art parallelization of SURF on GPGPU, our system achieves a speedup by a factor of 2.3X.

A High Performance Image Authentication Algorithm on GPU with CUDA

A CUDA Based Implementation of an Image Authentication Algorithm

Data Processing and Applied Technology in the Algorithm and Cracking of NTLM-Hash - A Highly Proficient Parallel Computing Architecture

An Efficient Acceleration of Symmetric Key Cryptography Using General Purpose Graphics Processing Unit

High Throughput Implementation of MD5 Algorithm on GPU

Accelerating High-resolution Image Stitching for the Dual Camera System Based on GPU

High-Performance Implementation of the Identity-Based Signature Scheme in IEEE P1363 on GPU

A GPU-paralleled implementation of an enhanced face recognition algorithm

A Program Behavior Study of Block Cryptography Algorithms on GPGPU

CUDA Based High Performance Implementation of RSA Algorithm

Research of Fast Mask Dodging for Satellite Image Based on CUDA

Research of Image Classification Algorithm Based on GPU

Effective and Efficient Image Copy Detection Based on GPU.

Design and Optimizations of the MD5 Crypt Cracking Algorithm Based on CUDA.

Implementation and Evaluation of Different Parallel Designs of AES Using CUDA

Fast Face Detection Algorithm Based on CUDA

CUDA Optimization Strategies for Compute- and Memory-Bound Neuroimaging Algorithms

A GPU-based high-throughput image retrieval algorithm

GPU-Based High Performance Password Recovery Technique for Hash Functions

Design and Performanc Optimization of Surveillance Video Inter-frame Forgery Detection System

The Implementation and Performance Analysis of AWMMF Parallel Algorithm on GPU