Table 7.
The summary of recent machine learning based colorimetric analysis.
| Learning type | Purpose | Image format | Machine learning model | Software | Training data | Color space | Sample | Ref |
|---|---|---|---|---|---|---|---|---|
| Supervised learning | Classification | RAW JPEG |
Least-Squares Support-Vector Machine (LS-SVM) | MATLAB | 385 images | RGB, HSV, LAB | Hydrogen peroxide | 208 |
| Supervised learning | Classification | JPEG | Linear Discriminant analysis (LDA), Gradient Boosting Classifier (GBC), Random forest RF) | Python, MATLAB | 224 images | RGB, HSV, LAB | Artificial Saliva | 94 |
| Supervised learning | Classification | JPEG | LDA, SVM, ANN | MATLAB, Python, Android studio | - | RGB, HSV, YUV, Lab | Alcohol solution | 209 |
| Supervised learning | Classification | JPEG | LDA Ensemble bagging classifier (EBC) |
Matlab, Android studio |
616 images | RGB, HSV, YUV, LAB | Artificial saliva | 210 |
| Supervised learning | Classification | JPEG | Convolutional neural network (CNN) | MATLAB | 1600 images | RGB | Glucose solution | 211 |
| Supervised learning | Classification | Spectrum | Support vector machine-radial basis function (SVM-RBF) | - | - | - | Glucose solution | 212 |
| Supervised learning | Classification | JPEG | Multi-Layer Perceptron (MLP), Residual Network (ResNet), CNN | - | 490 images | RGB | C-reactive protein (CRP) | 212 |
| Supervised learning | Classification | JPEG, RAW | LS-SVM | MATLAB | 450 images | RGB | prepared PH solution | 213 |
| Supervised learning | Classification | JPEG, Spectrum | Faster region-based (CNN) | - | 1500 images | RGB | Urine | 102 |
| Supervised learning | Classification | RAW | Artificial neural networks (ANNs) | MATLAB | 160 and 54 data points | CMYK | Artificial urine | 62 |
| Supervised learning | Classification | NIR Spectrum |
Deep neuronal network (DNN) | - | 1024 dataset | - | Serum glucose | 39 |
| Supervised learning | Classification | Spectrum | Multi-Channel -CNN | - | - | - | Glucose solution | 214 |