A morphological feature is a measurable physical characteristic describing an object's size, shape, or geometry. In raisin image analysis, these characteristics are represented by pixel measurements or quantities derived from pixel counts and can support variety classification. section 5.2
The raisin data set uses seven numeric morphological features:
- Area: the number of pixels within the raisin boundary. section 5.2.1
- MajorAxisLength: the longest line that can be drawn across the raisin grain. section 5.2.1
- MinorAxisLength: the shortest line that can be drawn across the raisin grain. section 5.2.1
- Eccentricity: how elongated or stretched the grain is compared with an ellipse. section 5.2.1
- ConvexArea: the number of pixels inside the area enclosed by the grain's outer perimeter. section 5.2.1
- Extent: the ratio of the grain's region to the total pixels in its bounding box. section 5.2.1
Raisin variety classification
A data set containing 450 Kecimen and 450 Besni raisins uses morphological features as numeric predictors and variety as a binary outcome. Grouped statistics, segmented histograms, and class-specific correlation heat maps help assess whether the measured characteristics distinguish the varieties before fitting logistic regression. section 5.2 section 5.2.1
>>> import pandas as pd
>>> raisins = pd.read_csv('raisin_dataset.csv')
Load the raisin measurements into a pandas DataFrame before inspecting feature definitions, distributions, missing values, and relationships. section 5.2.1
