Parallel and SIMD optimization of image feature extraction

Citations of this article
Mendeley users who have this article in their library.


Image feature extraction is widely used in content-based image retrieval(CBIR), computer version and all kinds of image processing applications. In this paper, we introduce some parallel and SIMD optimizations of image-feature extraction to overcome the disadvantages of original methods. We mainly use both thread-parallel optimization and Single Instruction Multiple Data (SIMD) optimizations. And especially, for some hot point, we use SIMD logical operations to eliminate the random conditional branches which cannot be effectively predicted by CPU Branch Prediction. We experimented our optimized implementation on multi-core systems, and various images were used to test the image-feature extraction results and performance. All the parallel and SIMD optimizations work out a good cumulative performance speedup. © 2011 Published by Elsevier Ltd.




Qi, M., Sun, G., & Chen, G. (2011). Parallel and SIMD optimization of image feature extraction. In Procedia Computer Science (Vol. 4, pp. 489–498).

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free