Scholarlink

This research explores the application of deep learning techniques to classify species and sex of Leptograpsus crabs. The dataset comprised of various morphological measurements, that included the carapace width and length. A convolutional neural network (CNNs) for feature extraction and classification was used. The study achieved 97.5% accuracy for species classification and 95% accuracy for sex classification, with high ROC AUC values (1.0 for species and 0.99 for sex). To ensure model robustness overfitting prevention strategies, such as early stopping and dropout layers, were employed. Additionally, SHAP analysis identified carapace width and carapace length as the most influential features. This research demonstrates the efficacy of advanced AI methodologies in biological classification, contributing to both artificial intelligence and ecological research, with potential applications in biodiversity monitoring and conservation efforts.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Leveraging Deep Learning for Accurate Classification of Leptograpsus Crabs Based on Morphological Measurements

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Usman Malik,
  • Mustafa Malik,
  • Asif Malik

摘要

This research explores the application of deep learning techniques to classify species and sex of Leptograpsus crabs. The dataset comprised of various morphological measurements, that included the carapace width and length. A convolutional neural network (CNNs) for feature extraction and classification was used. The study achieved 97.5% accuracy for species classification and 95% accuracy for sex classification, with high ROC AUC values (1.0 for species and 0.99 for sex). To ensure model robustness overfitting prevention strategies, such as early stopping and dropout layers, were employed. Additionally, SHAP analysis identified carapace width and carapace length as the most influential features. This research demonstrates the efficacy of advanced AI methodologies in biological classification, contributing to both artificial intelligence and ecological research, with potential applications in biodiversity monitoring and conservation efforts.