Table 5. Feature for image enhancement.
The diversity of blood feature selection used in recent studies. It also highlights the high accuracy percentages achieved by various classifiers and the effectiveness of the selected features in blood cell segmentation and classification tasks, which produces outstanding results.
| Work | Features | Classifier | Database | Accuracy % |
|---|---|---|---|---|
| Du et al. (2019) | Morphological feature | CNN | 17933 samples | 90.7 |
| Mishra, Majhi & Sa (2019) | DOST+PCA+LDA, | ADBRF, RF | ALL-IDB1 | 98.6, 99.6 |
| Liu & Long (2019) | Morphological feature | Augmented enhanced bagging classifier | 76 images | 85 |
| Hegde et al. (2019) | Morphological feature | TissueQuant algorithm | ALL-IDB1 | 96.5 |
| Sharma & Buksh (2019) | Morphological feature | fuzzy based CNN | ALL-IDB1 | 98.02 |
| Al-jaboriy et al. (2019) | 4-moment statistical features | ANN | ALL-IDB1 | 97 |
| Agaian, Madhukar & Chronopoulos (2018) | Cell energy, color features and morphological | SVM | ALL-IDB1 | 97 |
| Liang et al. (2018) | Size and intensity of the nucleus | CNN-RNN | BCCD dataset | 94.3 |
| Hegde et al. (2018) | Area, perimeter, circulatory, convexity and solidity | Hybrid-classifier (SVM & NN) | 117 images acquired by themselves. | 93.4 |
| Sampathila, Shet & Basu (2018) | GLCM | GUI | Leishman-stained thin blood smear slides | 96.7 |
| Kihm et al. (2018) | Shape, size | CNN | 4,000 manually classified images | 91.8 |
| Das, Maiti & Chakraborty (2018) | Mean intensity, standard deviation, skewness, kurtosis, and entropy | Random Forest (RF) |
950 nucleated blood cells | 99.42 |
| Mishra, Majhi & Sa (2018) | GLRLM and texture | SVM | ALL-IDB1 | 96.97 |
| Xu et al. (2017) | Shape | Deep CNN | 7,000 single RBCs images | 91.01 |
| Acharya & Kumar (2017) | Area, perimeter, diameter, shape and geometric | Modified watershed transform |
Nil | 98 |
| Mishra et al. (2017a) | GLCM and PCA (nucleus) | RF | ALL-IDB1 | 99.04 |
| Mishra et al. (2017b) | Discrete cosine transform (DCT) | SVM−L | ALL-IDB | 89.76 |
| Singhal & Singh (2016) | LBP and GLCM | SVM | ALL-IDB1 | 93.8 |
| Di Ruberto, Loddo & Putzu (2016) | Morphological, color, and textural | SVM−P | ALL-IDB1, ALL-IDB2 | 93.2 |
| Rawat et al. (2015) | Shape and texture features | ANN | ALL-IDB1 | 97.2 |
| Neoh et al. (2015) | Color, shape, and texture | Dempster-shafer | ALL-IDB1 | 96.7 |
| Mohapatra, Patra & Satpathy (2014) | Morphological, color, and texture | EOC5 | Leishman-stained peripheral blood sample |
99 |
| Nasir, Mashor & Hassan (2013) | Size, color, and shape | MLP | 500 manually collected images | 95.70 |