Abstract:
Indigenous medicinal plants hold immense therapeutic potential, yet their systematic
identification and medicinal assessment remain underexplored. This study addresses the
research gap by investigating the integration of traditional knowledge with modern technology
to identify and assess the medicinal properties of plant species in the Yem Special Zone,
Ethiopia. A novel Medicinal Assessment Framework (MAF) is developed using image processing
techniques and machine learning classification to bridge the gap between traditional practices
and contemporary scientific methodologies. The study employs a Design Science Research
Methodology (DSRM), involving iterative cycles of development, evaluation, and refinement. Key
steps include ethnobotanical data collection, digitization of traditional knowledge, acquisition
and preprocessing of plant leaf images, feature extraction (shape, color, and texture), machine
learning-based classification, and medicinal assessment scoring. Advanced image processing
techniques such as Gaussian blur, Otsu’s thresholding, and morphological operations are used,
while machine learning classifiers assess the extracted features. The results demonstrate the
efficacy of the framework, achieving satisfactory accuracy in classifying 25 indigenous plant
species based on leaf characteristics. The medicinal assessment scoring system combines
traditional knowledge with quantitative analysis, offering a comprehensive evaluation of each
species' medicinal potential. This integrative approach not only enhances the identification
process but also contributes to the preservation and sustainable use of indigenous medicinal
plants. In conclusion, the proposed framework provides a valuable tool for researchers,
conservationists, and healthcare practitioners, promoting interdisciplinary research and
community engagement. This work underscores the importance of leveraging traditional
knowledge alongside cutting-edge technologies to preserve cultural heritage and biodiversity
while advancing medicinal research.
