Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{185885,
        author = {Mr. Madhan H K and Mr. Anil Kumar K N},
        title = {DeepSeaVision: Transfer-Tuned CNN Ensembles for Accurate Marine Species Classification},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {5},
        pages = {3675-3682},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=185885},
        abstract = {DeepSeaVision is an AI system that automatically identifies underwater marine species from images. It serves biodiversity conservation and research by offering reliable, real-time classification. The web platform currently distinguishes Crabs, Jelly Fish, Seahorse, Sharks, Starfish, and Turtle/Tortoise. Data are organized into train, validation, and test splits, with ~200 - 300 high-resolution images per class. All images are resized to 256×256 RGB to standardize input. Convolutional Neural Networks handle feature extraction, capturing spatial and visual patterns. Four classifiers are used: a Custom CNN and fine-tuned EfficientNetB0, ResNet50, and InceptionV3. Models produce softmax probabilities for each class. InceptionV3 achieves the best performance, at roughly 94% accuracy on the project dataset. A React.js frontend supports drag-and-drop uploads and webcam capture for a responsive UX. A Flask or Node.js backend processes images and invokes TensorFlow/Keras models for predictions. When confidence is low, the system returns “Unknown Creature” to minimize misclassification and build trust.},
        keywords = {Deep learning, CNN, marine species classification, image recognition, real-time prediction, React.js, TensorFlow, Flask, InceptionV3.},
        month = {October},
        }