Analysis of a Step-Based Watershed Algorithm Using CUDA

Vitor, Giovani Bernardes; Körbes, André; Lotufo, Roberto de Alencar; Ferreira, Janito Vaqueiro

doi:10.4018/978-1-4666-1574-8.ch018

Cited by 4 publications

(7 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Körbes et al [26] introduced an iterative parallel watershed algorithm, where an image is divided into 16 × 16 tiles, and each tile computes a small local watershed transformation, for each iteration of the algorithm. Collins et al [25] mapped the co-segmentation problem to linear algebra operations, which were then solved with CUDA, offering a high-quality solution at a high computational cost.…”

Section: Related Workmentioning

confidence: 99%

“…Due to segmentation being a critical step in a large number of problems, there is now a vast amount of research on this subject, with many different approaches: clustering and dual clustering [1]- [3], histogram [11], edge detection [7], [11], region growing [3]- [5], graph partitioning [8]- [11], watershed [11], [26], adaptive thresholds [6], split and merge [7], [11], mean shift and mode seeking [10], [13], [24], [31], hierarchical [11], [15] and active contours [11], [17], [18].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

GPU Sparse Ray-Traced Segmentation

Petrescu¹,

Morar²,

Moldoveanu³

et al. 2019

IEEE Access

View full text Add to dashboard Cite

This paper introduces a real-time region growing segmentation algorithm, designed for graphics processing units (GPUs), which labels only a fraction of the input elements. Instead of searching locally around each element for strong similarity, like state-of-the-art segmentation and pre-segmentation methods do, the proposed algorithm searches both locally and remotely, using a unique ray tracing-based search strategy, which quickly covers the segmentation search space. The presented algorithm fully exploits the parallelism of the GPUs, sparsely segmenting high-resolution images (4K) in real-time on low range laptops and other mobile devices, approximately 5× times faster than the state-of-the-art simple linear iterative clustering (SLIC). While this paper demonstrates the results with images, the algorithm is trivially modifiable to work with input sets of any dimension. In contrast to the state-of-the-art real-time GPU methods, this algorithm doesn't require additional merging steps, as pre-segmentation methods do, and it produces complete segmentation. Additionally, post-segmentation optional stages for complete labeling and region merging on the GPU are also provided, although they are not always necessary.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

GPU Sparse Ray-Traced Segmentation

Petrescu¹,

Morar²,

Moldoveanu³

et al. 2019

IEEE Access

View full text Add to dashboard Cite

show abstract

“…The intuitive idea of the Watershed Transform comes from the geography: landscape of topography reliefs responsible for the formation of watersheds, these being divided into lines of domain aiming the attraction of rainwater over the region. An alternative approach is to imagine a landscape being immersed in a lake, with holes drilled in local minimums [5] [6].…”

Section: A Image Segmentationmentioning

confidence: 99%

“…As a result, the landscape is fractioned into regions or basins separated by dams called basin hydrography lines. The image segmentation process presents wide applicability, being used to solve problems in several areas [5] [6].…”

Section: A Image Segmentationmentioning

confidence: 99%

Proposal of a Medical Algorithm Based on the Application of Digital Image Processing and Visual Communication Techniques

Monteiro¹,

Iano²,

França³

2018

SET IJBE

View full text Add to dashboard Cite

Blood tests help detect various types of diseases, where the erythrocytes count (RBC) and leukocytes (WBC) can use automatic and/or manual methodologies. The evaluation of RBC and WBC has a direct impact on the diagnosis of anemia, leukemia, viral, parasitic and viral infections. Given the importance and wide applicability of the Watershed Transform (WT) and Morphological Operations (MO), they can segment medical images to increase the efficiency and reliability of medical diagnostics. This study uses a WT-MO algorithm for segmentation, detection, and counting of blood cells that attend the criteria of efficiency and reliability. This methodology may be the first step in making blood tests more accessible to people from developing and underdeveloped countries. The WT-MO algorithm has been benchmarked using 30 microscopy images of a blood smear. The WT-MO algorithm results presented high accuracy (93%). The simulations of the algorithm executed in different hardware platforms presented average simulation and processing time with less than 3 seconds per sample. Therefore, the WT-MO based algorithm is accurate, reliable and a low-cost technique, which can be applied as a third methodology to perform the laboratory tests and to speed up further medical diagnostic.

show abstract

“…More recent work suggests hybrid [15,10] or fully parallel watershed algorithms by topographical distance [16,17,9,12,13]. For each pixel in the image, a path of steepest descent is constructed and can be followed down to the nearest minimum.…”

Section: Introductionmentioning

confidence: 99%

Path Reducing Watershed for the GPU

Yeghiazaryan

Voiculescu

2018

2018 IEEE Winter Conference on Applications of Computer Vision (WACV)

View full text Add to dashboard Cite

The watershed transform is a popular image segmentation procedure from mathematical morphology used in many applications of computer vision. This paper proposes a novel parallel watershed procedure designed for GPU implementation. Our algorithm constructs paths of steepest descent and reduces these paths into direct pointers to catchment basin minima in logarithmic time, also crucially incorporating successful resolution of plateaux. Three implementation variants and their parameters are analysed through experiments on 2D and 3D images; a comparison against the state-of-the-art shows a runtime improvement of around 30%. For 3D images of 128 megavoxels execution times of approximately 1.5-2 seconds are achieved.

show abstract

Analysis of a Step-Based Watershed Algorithm Using CUDA

Cited by 4 publications

References 20 publications

GPU Sparse Ray-Traced Segmentation

GPU Sparse Ray-Traced Segmentation

Proposal of a Medical Algorithm Based on the Application of Digital Image Processing and Visual Communication Techniques

Path Reducing Watershed for the GPU

Contact Info

Product

Resources

About