GPU architecture evaluation for multispectral and hyperspectral image analysis

Fresse, Virginie; Houzet, Dominique; Gravier, Christophe

doi:10.1109/dasip.2010.5706255

Cited by 17 publications

(18 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to maintain the coalescence of memory accesses, we use a spatial-domain partitioning which divides the image into cubes of pixels and bands in a similar way as [3] for computing the initial block reduction and the aggregation of results at each iteration of the algorithm. For the ANNs computation steps, partitioning strategies similar to those described for the target detection algorithm at pixel level can be used.…”

Section: B Multiresolution Target Detection Algorithmmentioning

confidence: 99%

“…These parameters are stored in the global memory of the GPU. Then at each iteration these values are aggregated to compute for each window of interest the reductions over the whole window for obtaining (3). In total, 2mW H/(8 × 8) parameters have to be stored, where m is the number of spectral bands and W × H is the size of the image.…”

Section: B Multiresolution Target Detection Algorithmmentioning

confidence: 99%

“…Applications range from conventional remote sensing monitoring and exploration of the landscape, to manufacturing plant material surface inspection, art authentication or search and rescue operations among others [1]- [3]. A hyperspectral image is basically an image where for each pixel we have a set of m values that correspond to the spectral components in the visible or infrared areas as shown in Fig.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Towards real-time hyperspectral image processing, a GP-GPU implementation of target identification

Heras

Argüello

Gómez

et al. 2011

Proceedings of the 6th IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems

View full text Add to dashboard Cite

In the quest for real time processing of hyperspectral images, this paper presents two artificial intelligence algorithms for target detection specially developed for their implementation over GPU and applied to a search-andrescue scenario. Both algorithms are based on the application of artificial neural networks to the hyperspectral data. In the first algorithm the neural networks are applied at the level of individual pixels of the image. The second algorithm is a multiresolution based approach to scale invariant target identification using a hierarchical artificial neural network architecture. We have studied the main issues for the efficient implementation of the algorithms in GPU: the exploitation of thousands of threads that are available in this architecture and the adequate use of bandwidth of the device. The tests we have performed show both the effectiveness of detection of the algorithms and the efficiency of the GPU implementation in terms of execution times.

show abstract

Section: B Multiresolution Target Detection Algorithmmentioning

confidence: 99%

Section: B Multiresolution Target Detection Algorithmmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Towards real-time hyperspectral image processing, a GP-GPU implementation of target identification

Heras

Argüello

Gómez

et al. 2011

Proceedings of the 6th IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems

View full text Add to dashboard Cite

show abstract

“…The main differences between them in the above mentioned order are with respect to a greater number of wavelength bands and higher precision. 1,10 Both criteria use the number of wavelength bands as one of the defining parameters. 1,10 Both criteria use the number of wavelength bands as one of the defining parameters.…”

Section: Introductionmentioning

confidence: 99%

Pushbroom hyperspectral imaging system with selectable region of interest for medical imaging

Lim

Murukeshan

2015

J. Biomed. Opt

View full text Add to dashboard Cite

A spatial-scanning pushbroom hyperspectral imaging (HSI) system incorporating a video camera (VC) which is not only used for direct video imaging but also for the selection of the region of interest within the VC’s full field-of-view is presented. Using a VC for these two applications brings many benefits to a pushbroom HSI system, such as a minimized data acquisition time and smaller data storage requirement. A detailed description of the system followed by the methods and formulas used for calibration and electronic hardware interfacing were discussed and analyzed using United States Air Force resolution chart, chicken breast tissue, and fluorescent targets as test samples. The proposed concepts and developed system can find potential biomedical imaging applications and can be extended to endoscopic imaging applications as well.

show abstract

“…With an in-depth study and several necessary improvements on typical working principles of graphics processor [1], we develop a graphics processing pipeline organization tree [2] as is shown in Figure.1. A data pipeline organization of this kind using the dual-rail handshake protocol [3] makes our simulation designing platform much closer to that of a real time graphics processor [4].…”

Section: Introductionmentioning

confidence: 99%

A Study on the Algorithms for 2D Texture Mapping and Its Employment in Hardware Design

Wang¹,

Dong²,

Du³

2012

Proceedings of the 2nd International Conference on Computer Application and System Modeling

View full text Add to dashboard Cite

Abstract-Based on a solid analysis of the working principle of graphic processors, this paper puts forward the dual-rail handshake protocol of the data pipeline which is to be proved to improve the architecture of graphics rendering pipeline and thus make the system much closer to that of a graphic processor in real time. Specific algorithms and hardware design methods are as well developed in this work with their focus dwelling on the computation-consuming texture mappings, including two critical techniques: texture coordinate calculation and texture filtering. FPGA result shows that circuits based on textual algorithms can define accurately the textual coordinates and yield precise textual images, bringing significant improvements to the performance of graphics rendering.

show abstract

GPU architecture evaluation for multispectral and hyperspectral image analysis

Cited by 17 publications

References 8 publications

Towards real-time hyperspectral image processing, a GP-GPU implementation of target identification

Towards real-time hyperspectral image processing, a GP-GPU implementation of target identification

Pushbroom hyperspectral imaging system with selectable region of interest for medical imaging

A Study on the Algorithms for 2D Texture Mapping and Its Employment in Hardware Design

Contact Info

Product

Resources

About