Designing efficient SIMD algorithms for direct Connected Component Labeling

Abstract : Connected Component Labeling (CCL) is a fundamental algorithm in computer vision, and is often required for real-time applications. It consists in assigning a unique number to each connected component of a binary image. In recent years, we have seen the emergence of direct parallel algorithms on multicore CPUs, GPUs and FPGAs whereas, there are only iterative algorithms for SIMD implementation. In this article, we introduce new direct SIMD algorithm for Connected Component Labeling. They are based on the new Scatter-Gather, Collision Detection (CD) and Vector Length (VL) instructions available in the recent Intel AVX512 instruction set. These algorithms have also been adapted for multicore CPU architectures and tested for each available SIMD vector length. These new algorithms based on SIMD Union-Find algorithms can be applied to other domains such as graphs algorithms manipulating Union-Find structures.
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-02049029
Contributor : Arthur Hennequin <>
Submitted on : Tuesday, May 7, 2019 - 11:27:00 AM
Last modification on : Wednesday, May 15, 2019 - 9:43:26 AM

Identifiers

Citation

Arthur Hennequin, Ian Masliah, Lionel Lacassagne. Designing efficient SIMD algorithms for direct Connected Component Labeling. WPMVP'19 - 5th Workshop on Programming Models for SIMD/Vector Processing, Feb 2019, Washington, United States. pp.4:1--4:8, ⟨10.1145/3303117.3306164⟩. ⟨hal-02049029⟩

Share

Metrics

Record views

13

Files downloads

13