Semi-supervised classification of microorganism
A system and method that identify and classify unknown microorganisms and/or known microorganisms with anomalies are provided. The system and method comprise processing images of microorganisms from an aquatic environment; extracting features from the processed images; an unsupervised partitioning algorithm for identifying and classifying known microorganisms in the aquatic environment based upon the extracted features; and a supervised classifier neural network that is trained with the unsupervised partitioning algorithm and identifies and classifies unknown microorganisms and/or known microorganisms with anomalies.
1. A method comprising:
classifying known species from among a population of microorganisms, wherein each of the known species is classified according to a collection of features extracted from image data;
developing a neural network for each of the known species;
applying each of the neural networks to the population of microorganisms, thereby identifying microorganisms with features that are different from the features of the known species;
identifying (i) known species with anomalies from within the population of microorganisms and (ii) unknown species, based upon the features that are different from the features of each of the known species.
2. The method of claim 1 , wherein the collection of features is selected from the group consisting of shape, size, texture, structure, behavior, and combinations thereof.
3. The method of claim 1 , wherein the known species are classified with an unsupervised partitioning algorithm.
4. The method of claim 3 , wherein the unsupervised partitioning algorithm is selected from the group consisting of a partition entropy algorithm, a purity algorithm, a random forest algorithm, a clustering algorithm, and combinations thereof.
5. The method of claim 4 , wherein the clustering algorithm is selected from k-Means, Fuzzy k-Means, Gaussian Mixture Model (GMM), and combinations thereof.
6. The method of claim 3 , wherein the neural network is trained with the unsupervised partitioning algorithm.
7. The method of claim 1 , wherein the neural network is selected from the group consisting of an artificial neural network, a convolution neural network, a random forest algorithm, and combinations thereof.
8. The method of claim 1 , wherein the microorganisms are selected from the group consisting of plankton, flagella, amoeba, paramecia, bacteria, protozoans, eukaryotic organelles, prokaryotic organelles, and combinations thereof.
9. A method comprising:
classifying known plankton species from among a population of different plankton species, wherein each of the known plankton species is classified according to a collection of features extracted from image data;
developing a neural network for each of the known plankton species;
applying each neural network to the population of different plankton species, thereby identifying plankton species with features that are different from the features of the known plankton species; and
identifying (i) known plankton species with anomalies from within the population of different plankton species and (ii) unknown plankton species, based upon the features that are different from the features of the known plankton species.
10. The method of claim 9 , wherein the collection of features is selected from the group consisting of shape, size, texture, structure, behavior, and combinations thereof.
11. The method of claim 9 , wherein the known species are classified with an unsupervised partitioning algorithm.
12. The method of claim 11 , wherein the neural network is trained with the unsupervised partitioning algorithm.
13. The method of claim 11 , wherein the unsupervised partitioning algorithm is selected from the group consisting of a partition entropy algorithm, a purity algorithm, a random forest algorithm, a clustering algorithm, and combinations thereof.
14. The method of claim 13 , wherein the clustering algorithm is selected from k-Means, Fuzzy k-Means, Gaussian Mixture Model (GMM), and combinations thereof.
15. The method of claim 9 , wherein the neural network is selected from the group consisting of an artificial neural network, a convolution neural network, a random forest algorithm, and combinations thereof.