摘要
Two additional features are particularly useful in pixelwise satellite data segmentation using neural networks: one results from local window averaging around each pixel (MWA) and another uses a standard deviation estimator (MWSD) instead of the average. While the former’s complexity has already been solved to a satisfying minimum, the latter did not. This article proposes a new algorithm that can substitute a <i><span style="font-family:Verdana;">naive</span></i><span style="font-family:Verdana;"> MWSD, by making the complexi</span><span><span style="font-family:Verdana;">ty of the computational process fall from </span><i><span style="font-family:Verdana;">O</span></i><span style="font-family:Verdana;">(</span><i><span style="font-family:Verdana;">N</span></i><sup><span style="font-family:Verdana;">2</span></sup><i><span style="font-family:Verdana;">n</span></i><sup><span style="font-family:Verdana;">2</span></sup><span style="font-family:Verdana;">) to </span><i><span style="font-family:Verdana;">O</span></i><span><span style="font-family:Verdana;">(</span><i><span style="font-family:Verdana;">N</span></i></span><sup><span style="font-family:Verdana;">2</span></sup><i><span style="font-family:Verdana;">n</span></i><span style="font-family:Verdana;">)</span><span style="font-family:Verdana;">, where </span><i><span style="font-family:Verdana;">N</span></i><span style="font-family:Verdana;"> is a square</span></span><span style="font-family:Verdana;"> input array side, and </span><i><span style="font-family:Verdana;">n</span></i><span style="font-family:Verdana;"> is the moving window’s side length. The Num</span><span style="font-family:Verdana;">ba python compiler was used to make python a competitive high-performance</span> <span style="font-family:Verdana;">computing language in our optimizations. Our results show efficiency benchmars</span>
Two additional features are particularly useful in pixelwise satellite data segmentation using neural networks: one results from local window averaging around each pixel (MWA) and another uses a standard deviation estimator (MWSD) instead of the average. While the former’s complexity has already been solved to a satisfying minimum, the latter did not. This article proposes a new algorithm that can substitute a <i><span style="font-family:Verdana;">naive</span></i><span style="font-family:Verdana;"> MWSD, by making the complexi</span><span><span style="font-family:Verdana;">ty of the computational process fall from </span><i><span style="font-family:Verdana;">O</span></i><span style="font-family:Verdana;">(</span><i><span style="font-family:Verdana;">N</span></i><sup><span style="font-family:Verdana;">2</span></sup><i><span style="font-family:Verdana;">n</span></i><sup><span style="font-family:Verdana;">2</span></sup><span style="font-family:Verdana;">) to </span><i><span style="font-family:Verdana;">O</span></i><span><span style="font-family:Verdana;">(</span><i><span style="font-family:Verdana;">N</span></i></span><sup><span style="font-family:Verdana;">2</span></sup><i><span style="font-family:Verdana;">n</span></i><span style="font-family:Verdana;">)</span><span style="font-family:Verdana;">, where </span><i><span style="font-family:Verdana;">N</span></i><span style="font-family:Verdana;"> is a square</span></span><span style="font-family:Verdana;"> input array side, and </span><i><span style="font-family:Verdana;">n</span></i><span style="font-family:Verdana;"> is the moving window’s side length. The Num</span><span style="font-family:Verdana;">ba python compiler was used to make python a competitive high-performance</span> <span style="font-family:Verdana;">computing language in our optimizations. Our results show efficiency benchmars</span>
作者
Marcos R. de A. Conceição
Luis F. F. de Mendonça
Carlos A. D. Lentini
Marcos R. de A. Conceição;Luis F. F. de Mendonça;Carlos A. D. Lentini(Oceanography Department, Geoscience Institute of the Federal University of Bahia (UFBA), Salvador, Brazil;Earth and Environmental Physics Department, Physics Institute of the Federal University of Bahia (UFBA), Salvador, Brazil;Tropical Oceanography Group (GOAT), Salvador, Brazil)
