Table 15 Complexity analysis of key Modules.

From: Collaborative representation and confidence-driven semi-supervised learning for hyperspectral image classification

Module

Time Complexity

Space Complexity

Graph Construction

\(\:O\left(N\right)\)

\(\:O\left(N\right)\)

GCN Training

\(\:O(T\times\:NCH)\)

\(\:O(T\times\:NCH)\)

Region Evaluation

\(\:O(R\times\:{S}^{3})\)

\(\:O(NH+CH+{H}^{2}+HK)\)

Adaptive Subdivision

\(\:O(R{\prime\:}\times\:SKI)\)

\(\:O\left(R{\prime\:}S\right)\)

Dynamic Ensemble Selection

\(\:O(R\times\:M)\)

\(\:O\left(R\right)\)

Result Integration

\(\:O(N\times\:HW)\)

\(\:O\left(HW\right)\)

  1. Notations: \(\:N\): valid pixels;\(\:\:C,H,K\): input/hidden/class dimensions;\(\:\:R,R{\prime\:}\): total/low-performance regions; \(\:S\): average region size; \(\:M\): classifier count; \(\:I\): K-means iterations.
Back to article page