How to speed Connected Component Labeling up with SIMD RLE algorithms

Lemaître, F.; Hennequin, A. M.; Lacassagne, Lionel

doi:10.1145/3380479.3380481

Cited by 8 publications

(14 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Literature on CCL algorithms is extensive and has been centered on 2D images. CCL on CPUs has been heavily studied and optimized [14][17][6] [26]. On GPUs, after an early era of iterative algorithms [43][3] [20], a new generation introduced by Komura [23] are now direct; a new way to manage equivalences and reduce memory accesses was introduced by Playne [36] and has become the basis of the fastest CCL algorithms [19] [2].…”

Section: State-of-the-art Of 3d Algorithmsmentioning

confidence: 99%

“…Overlapping segments between lines can also be found without ER using a Finite-State Machine (FSM). In the 2D unification [27], each state of the 2D FSM encodes segment configurations between the current and previous lines. Merging two lines involves iterating over both at the same time: a new label is created for each isolated segment, whereas the components of two overlapping segments are merged together.…”

Section: A Finite-state Machine-based Unificationmentioning

confidence: 99%

“…These results make the double-line algorithm at least as fast as the best algorithm on both OASIS and mitochondria (Figure 3). Fortunately, these steps lend themselves well to instruction level parallelism with SIMD [27]. Several SIMD implementations of the RLE and relabeling have been tested: SSE4, AVX2 and AVX512.…”

Section: Computational Reuse Of Merged Linesmentioning

confidence: 99%

See 2 more Smart Citations

LSL3D: A Run-Based Connected Component Labeling Algorithm for 3D Volumes

Maurice

Lemaître

Julien

et al. 2022

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Connect Component Labeling (CCL) has been a fundamental operation in Computer Vision for decades. Most of the literature deals with 2D algorithms for applications like video surveillance or autonomous driving. Nonetheless, the need for 3D algorithms is rising, notably for medical imaging. While 2D CCL algorithms already generate large amounts of memory accesses and comparisons, 3D ones are even worse. This is the curse of dimensionality. Designing an efficient algorithm should address this problem. This paper introduces a segment-based algorithm for 3D labeling that uses a new strategy to accelerate label equivalence processing to mitigate the impact of higher dimensions. We claim that this new algorithm outperforms State-of-the-Art algorithms by a factor from ×1.5 up to ×3.1 for usual medical datasets and random images.

show abstract

Section: State-of-the-art Of 3d Algorithmsmentioning

confidence: 99%

Section: A Finite-state Machine-based Unificationmentioning

confidence: 99%

See 1 more Smart Citation

LSL3D: A Run-Based Connected Component Labeling Algorithm for 3D Volumes

Maurice

Lemaître

Julien

et al. 2022

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

show abstract

“…There are already CPU algorithms implementing those ideas: the LSL [29] and derivatives. We re-designed FLSL [18], a variant of LSL for SIMD CPU (SSE, AVX512, Neon), to target GPUs and address their architectural constraints. The crucial part is to first do a segment detection that consists in an RLE encoder and relies on "compress-store" (Figure 2).…”

Section: Full Runs (Flsl)mentioning

confidence: 99%

“…CCL on CPUs has been heavily studied and optimized [15] [16] [17] [18]. Early GPU CCL algorithms were iteratives [19] [20] [21].…”

Section: Introductionmentioning

confidence: 99%

Taming Voting Algorithms on Gpus for an Efficient Connected Component Analysis Algorithm

Lemaître¹,

Hennequin

Lacassagne³

2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Self Cite

View full text Add to dashboard Cite

Connected Component Analysis is vastly used as a building block for many Computer Vision algorithms from many fields like medical image processing, surveillance, or autonomous driving. It extends Connected Component Labeling by computing some features of the connected components like their bounding box or their surface. As such, Connected Component Analysis is a voting algorithm just like histogram computation or Hough transform. Voting algorithms are difficult on many-core architectures like GPUs because of the serialization of atomic memory accesses. The trend to increase the number of cores makes this issue even more critical.This paper explores multiple ways to reduce those conflicts for voting algorithms and especially for Connected Component Analysis. We show that our new algorithm is from 4 up to 10 times faster than State-of-the-Art on average on an Nvidia A100.

show abstract

An Efficient Run-Based Connected Component Labeling Algorithm for Processing Holes

Lemaître¹,

Maurice²,

Lacassagne³

2022

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

This article introduces a new connected component labeling and analysis algorithm framework that is able to compute in one pass the foreground and the background labels as well as the adjacency tree. The computation of features (bounding boxes, first statistical moments, Euler number) is done on-the-fly. The transitive closure enables an efficient hole processing that can be filled while their features are merged with the surrounding connected component without the need to rescan the image. A comparison with State-of-the-Art shows that this new algorithm can do all these computations faster than all existing algorithms processing foreground and background connected components or holes.

show abstract

How to speed Connected Component Labeling up with SIMD RLE algorithms

Cited by 8 publications

References 27 publications

LSL3D: A Run-Based Connected Component Labeling Algorithm for 3D Volumes

LSL3D: A Run-Based Connected Component Labeling Algorithm for 3D Volumes

Taming Voting Algorithms on Gpus for an Efficient Connected Component Analysis Algorithm

An Efficient Run-Based Connected Component Labeling Algorithm for Processing Holes

Contact Info

Product

Resources

About