Related Concept Videos
Light Acquisition
Fruit Development, Structure, and Function
Plant Breeding and Biotechnology
Methods of Classification and Identification
You might also read
Related Articles
Articles linked to this work by shared authors, journal, and citation graph.
Assessment of the Potential of Bitter Melon (<i>Momordica charantia</i>) and Squash (<i>Cucurbita pepo</i>) as Rootstocks for Enhancing Drought Tolerance in Cucumber.
Breaking the silence on late-life depression: uncovering the drivers of help-seeking in Türkiye's ageing population.
Proteins from Edible Mushrooms: Nutritional Role and Contribution to Well-Being.
Identification and Classification of Snack-Type Watermelon (Citrullus lanatus) Genotypes Using Seed Morphology and
Uğur Ercan1, Sıtkı Ermiş2, Onder Kabas3
1Department of Informatics, Akdeniz University, 07070 Antalya, Türkiye.
Machine learning accurately identifies watermelon genotypes from seeds using morphological and color data. The Random Forest model achieved 92.22% accuracy, offering reliable genotypic classification for breeding and seed production.
More Related Videos
07:22Generating Homo- and Heterografts Between Watermelon and Bottle Gourd for the Study of Cold-responsive MicroRNAs
Published on: November 20, 2018
11:48A Contrast of Three Inoculation Techniques used to Determine the Race of Unknown Fusarium oxysporum f.sp. niveum Isolates
Published on: October 28, 2021
Area of Science:
- Agricultural Science
- Computer Science
- Genetics
Background:
- Accurate genotypic identification of watermelon (Citrullus lanatus) seeds is crucial for breeding programs and commercial production.
- Traditional methods for seed identification can be time-consuming and may lack precision.
- Developing automated, non-destructive methods for genotypic classification is highly desirable.
Purpose of the Study:
- To evaluate the effectiveness of machine learning (ML) algorithms for automatic identification of watermelon genotypes based on seed characteristics.
- To compare the performance of Artificial Neural Network (ANN), Random Forest (RF), and Extra Tree (ET) models in classifying watermelon seeds.
- To determine if morphological, physical, and colorimetric seed attributes can reliably predict genotype.
Main Methods:
- Collected data from nine watermelon genotypes, analyzing 200 seeds per genotype.
- Extracted morphological (size, shape), physical (weight), and colorimetric (L, a, b) attributes using high-resolution imaging and digital measurement.
- Trained and validated ANN, RF, and ET models using a 10-fold cross-validation approach.
Main Results:
- The Random Forest (RF) model demonstrated the highest performance, achieving 92.22% accuracy, 91.87% F1-score, and 0.9118 Cohen's Kappa.
- Extra Tree (ET) and Artificial Neural Network (ANN) models showed lower accuracy at 90.00% and 86.11%, respectively.
- Statistical analysis confirmed RF and ET significantly outperformed ANN, with RF offering superior stability.
Conclusions:
- Machine learning frameworks provide a rapid, reliable, and non-destructive method for classifying snack-type watermelon seeds by genotype.
- These ML approaches have strong potential for improving varietal traceability in breeding, quality control in seed production, and meeting industrial demands.
- The study validates the use of seed morphological, physical, and colorimetric data for accurate genotypic classification.
