Image retrieve for dolphins and whales based on EfficientNet network
Tong Zhou, Shengyang Li, Bo Tian, Yeming Cai · 2023
Image retrieve model for whales and dolphins, based on deep learning algorithms, analyzes features such as body shape, size, color, and tails to accurately classify images of different species. The technology has potential applications for scientific research, conservation, and marine wildlife management. It can help authorities identify species at risk of ship collision and reduce entanglement in fishing gear. Deep learning relies on large datasets to develop effective image retrieval models. We propose a network detects dorsal fins of whales and dolphins via feature extraction and image retrieval techniques. Architecture allows for efficient identification and outperforms comparison approaches like ResNet50 and EfficientNet B3, as evaluated on a dataset.