Table 5 Analysis of the machine learning algorithms.
From: Analysis of banana plant health using machine learning techniques
Process | Algorithm | Strength | Weakness |
|---|---|---|---|
Segmentation | K-means | Computationally very efficient and fast | Restricted mostly for spherical shape clustering |
Thresholding | Simplest image segmentation method | Difficult to find the threshold value | |
Feature extraction | HOG | No data loss as it directly works on the cell | Does not work for image rotation |
SIFT | Scale and rotation variant | A slow process | |
Classification | SVM | Works for linearly and nonlinearly separable classes | Supports only binary classification |
CNN | Works for different features instead of each pixel | Cannot handle rotation of the image |
