Diferenzas

Isto amosa as diferenzas entre a revisión seleccionada e a versión actual da páxina.

Enlazar a esta vista de comparación

hiperespectral:gpu-elm-rs [2016/05/12 13:26] (actual)
jorge.suarez creado
Liña 1: Liña 1:
 +====== GPU-Based Extreme Learning Machine for the Classification of Hyperspectral Remote Sensing Images ======
 +Experimental results related to the paper [[http://​ieeexplore.ieee.org/​xpl/​articleDetails.jsp?​arnumber=7005455&​sortType%3Dasc_p_Sequence%26filter%3DAND%28p_Publication_Number%3A4609443%29%26pageNumber%3D5%26rowsPerPage%3D50 |Efficient ELM-based Techniques for the Classification of Hyperspectral Remote Sensing Images on Commodity GPUs]] by Javier López-Fandiño,​ Pablo Quesada-Barriuso,​ Dora B. Heras, and Francisco Argüello, published in the [[http://​www.grss-ieee.org/​publications/​jstars/​|IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing]].
  
 +===== Abstract =====
 +Extreme Learning Machine (ELM) is an efficient learning algorithm that has been recently applied to hyperspectral image classification. In this paper, the first implementation of the ELM algorithm fully developed for Graphical Processing Unit (GPU) is presented. ELM can be expressed in terms of matrix operations so as to take advantage of the Single Instruction Multiple Data (SIMD) computing paradigm of the GPU architecture. Additionally,​ several techniques like the use of ensembles, a spatial regularization algorithm, and a spectralspatial classification scheme are applied and projected to GPU in order to improve the accuracy results of the ELM classifier. In
 +the last case, the spatial processing is based on the segmentation of the hyperspectral image through a watershed transform. The experiments are performed on remote sensing data for land cover applications achieving competitive accuracy results compared to analogous SVM strategies with significantly lower execution times. The best accuracy results are obtained with the spectralspatial scheme based on applying watershed and a spatially regularized ELM.
 +
 +===== Downloads =====
 +
 +== Hyperspectral images ==
 +  * Hypersectral datasets from Universidad del Pais Vasco: [[http://​www.ehu.es/​ccwintco/​index.php?​title=Hyperspectral_Remote_Sensing_Scenes|www.ehu.es]]
 +
 +== Execution Outputs ==
 +^   ​Indian Pines Dataset ​ ^^^^^^
 +|  Ground Truth          |  RCMG Output ​       |  Watershed ​ Output ​     |  ELM  Classification Map        |  Regularized ELM  Classification Map     ​| ​ Spectral-Spatial Classification Map | 
 +|{{:​inv:​downloadable_results:​gpu-elm-rs:​groundtruth_indianpines.png?​direct&​200| Ground Truth}}| {{:​inv:​downloadable_results:​gpu-elm-rs:​rcmg_indianpines_con8.png?​direct&​200|RCMG}}|{{:​inv:​downloadable_results:​gpu-elm-rs:​ca-watershed_indianpines_con8.png?​direct&​200|Watershed}} |{{:​inv:​downloadable_results:​gpu-elm-rs:​elm_classification_indianpines.png?​direct&​200|ELM}} | {{:​inv:​downloadable_results:​gpu-elm-rs:​regularized_elm_indianpines.png?​direct&​200|Regularized ELM}}| {{:​inv:​downloadable_results:​gpu-elm-rs:​spectral-spatial_classification_indianpines.png?​direct&​200|Spectral-Spatial Classification}}|
 +
 +^   Pavia Univ. Dataset ​   ^^^^^^
 +|  Ground Truth          |  RCMG Output ​       |  Watershed ​ Output ​     |  ELM  Classification Map        |  Regularized ELM  Classification Map     ​| ​ Spectral-Spatial Classification Map | 
 +|{{:​inv:​downloadable_results:​gpu-elm-rs:​groundtruth_paviauniv.png?​direct&​200| Ground Truth}}| {{:​inv:​downloadable_results:​gpu-elm-rs:​rcmg_paviauniv_con8.png?​direct&​200|RCMG}}|{{:​inv:​downloadable_results:​gpu-elm-rs:​ca-watershed_paviauniv_con8.png?​direct&​200|Watershed}} |{{:​inv:​downloadable_results:​gpu-elm-rs:​elm_classification_paviauniv.png?​direct&​200|ELM}} | {{:​inv:​downloadable_results:​gpu-elm-rs:​regularized_elm_paviauniv.png?​direct&​200|Regularized ELM}}| {{:​inv:​downloadable_results:​gpu-elm-rs:​spectral-spatial_classification_paviauniv.png?​direct&​200|Spectral-Spatial Classification}}|
 +
 +^   ​Salinas Dataset ​ ^^^^^^
 +|  Ground Truth          |  RCMG Output ​       |  Watershed ​ Output ​     |  ELM  Classification Map        |  Regularized ELM  Classification Map     ​| ​ Spectral-Spatial Classification Map |  ​
 +|{{:​inv:​downloadable_results:​gpu-elm-rs:​groundtruth_salinas.png?​direct&​200| Ground Truth}}| {{:​inv:​downloadable_results:​gpu-elm-rs:​rcmg_salinas_con8.png?​direct&​200|RCMG}}|{{:​inv:​downloadable_results:​gpu-elm-rs:​ca-watershed_salinas_con8.png?​direct&​200|Watershed}} |{{:​inv:​downloadable_results:​gpu-elm-rs:​elm_classification_salinas.png?​direct&​200|ELM}} | {{:​inv:​downloadable_results:​gpu-elm-rs:​regularized_elm_salinas.png?​direct&​200|Regularized ELM}}| {{:​inv:​downloadable_results:​gpu-elm-rs:​spectral-spatial_classification_salinas.png?​direct&​200|Spectral-Spatial Classification}}|
 +
 +
 +
 +
 +== Additional Results ==
 +{{:​inv:​downloadable_results:​gpu-elm-rs:​additionalresults-gpu-elm-rs.pdf|}} - Document containing results of additional experiments.
 +
 +
 +===== License =====
 +
 +:​cc-by-nc-nd:  ​
  • hiperespectral/gpu-elm-rs.txt
  • Última modificación: 2016/05/12 13:26
  • por jorge.suarez