Cross-modal learning for plankton recognition

Dimensions overall score 8.0

• Self-supervised cross-modal coordination can utilize multiple modalities and large volumes of unlabeled plankton data for plankton recognition.method 95%
• The proposed method achieves high recognition accuracy while requiring only a minimal number of labeled images.result 90%
• The cross-modal learning approach outperforms an image-only self-supervised baseline for plankton recognition.result 90%
• The method trains encoders for both image and profile modalities using binary supervisory information indicating same or different particle origin.method 95%
• Plankton recognition is achieved by employing a small labeled gallery of known plankton species combined with a k-NN classifier.method 95%
• The proposed method is inherently multimodal, capable of utilizing information from both image and profile data.technical 95%
• The approach is limited by dependence on specific instrument types that provide compatible optical profile data.limitation 80%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card