Fine-Tuned ResNet50 for Image-Based Sex Classification of Blue Swimming Crabs (Portunus Pelagicus)

Main Article Content

Micah Vivien Deocampo
Mark Dennis V. Manangan
Christian John Moncera

Abstract

Accurate sex identification of the blue swimming crab (Portunus Pelagicus) is essential for population monitoring, reproductive stock conservation, and sustainable fisheries management. Conventional morphological examination, however, requires trained personnel and may be affected by specimen orientation, image quality, and environmental conditions. This study evaluated a fine-tuned ResNet50 convolutional neural network for automated sex classification based on morphological image features. An experimental deep-learning approach was applied to a balanced secondary dataset comprising 4,180 labeled images, with equal representation of male and female crabs. The dataset was partitioned into 2,920 training, 620 validation, and 640 testing images. Transfer learning was used to adapt the pretrained ResNet50 architecture to binary classification. Image augmentation techniques, including rotation, flipping, brightness adjustment, and contrast modification, were also applied to increase feature diversity and reduce sensitivity to imaging variations. Model performance was assessed using classification accuracy and F1-score. On a hold-out evaluation set of 630 images, the best-performing ResNet50 model achieved a reported accuracy of 96.8% and an F1-score of 0.967, indicating promising classification performance. These findings demonstrate the potential of transfer learning to identify sex-related morphological characteristics in commercially important crustaceans. The proposed approach could support more consistent specimen screening, facilitate fisheries monitoring, and contribute to evidence-informed conservation practices. Further validation using independently collected field images is recommended to establish the model’s generalizability under operational conditions.

views Abstract Views: 0 times | download Downloaded: 0 times

Downloads

Download data is not yet available.

Article Details

How to Cite
Deocampo, M. V., Manangan, M. D. V., & Moncera, C. J. (2026). Fine-Tuned ResNet50 for Image-Based Sex Classification of Blue Swimming Crabs (Portunus Pelagicus). Research Journal of Education, Science and Technology, 5(1). https://doi.org/10.63179/rjest.v5i1.94
Section
Articles