Figure 7
From: Sugariness prediction of Syzygium samarangense using convolutional learning of hyperspectral images
Figure (a) presents the proposed 2-D convolutional layer for extracting spectral features of the HSIs H with size \(\left(W,L,\Lambda \right)\). Figure (b) shows the architecture of the proposed convolutional network for sugariness prediction. Input X is a 3-D HSIs cube sample, and the network contains four convolutional layers with a ReLU activation function. Batch normalization (BN) layers and Dropout (DP) layers are used to prevent overfitting. In the last Maxpooling layer, its output is a 1-D array obtained by calculating the maximum from each \(1\times 2\) patch of the feature map then flattening. The 1-D array was then fed to fully connected layers with two dense layers with BN and DP following to regress the sugariness prediction result Y.
