A 3D Fluorescence Classification and Component Prediction Method Based on VGG Convolutional Neural Network and PARAFAC Analysis Method

Three-dimensional fluorescence is currently studied by methods such as parallel factor analysis (PARAFAC), regional integration (FRI), and principal component (PCA). There are also many studies combining convolutional neural networks at present, but there no one method recognized the most effective among 3D analysis. Based on this, we took some samples from actual environment for measuring data obtained a batch of public datasets internet species. Firstly, preprocessed (including two steps PARAFAC CNN dataset generation), then proposed classification components fitting based VGG16 VGG11 networks. The network used with training accuracy 99.6% (as same PCA + SVM (99.6%)). Among maps networks, comprehensively compared improved LeNet network, AlexNet finally selected network. In training, MSE loss function cosine similarity to judge merit model, reached 4.6 × 10−4 (characterizing variability results results), criterion, 0.99 (comparison results). performance excellent. experiments demonstrate that has great application in