Content-Based Feature Extraction: Morphological Operators
Rik Kamal Kumar Das · 2020
The shape of an image is crucial to determine efficient feature vectors for content-based image classification (CBIC). Diverse standard operations like dilation and erosion are considered standard morphological operators to determine the shape of an object within an image. This property can be well utilized to extract meaningful features necessary to ensure higher accuracy of content-based image classification (CBIC). This chapter has demonstrated two different techniques—namely, top-hat transform and bottom-hat transform —for extraction of shape-based feature vectors from image data. The implementation guide is also provided in MATLAB . Extracted features with each of these techniques are compared against each other using various classification metrics and are finally evaluated for statistical significance of contribution toward enhanced classification performances.