Tighnari: Multi-modal Plant Species Prediction Based on Hierarchical Cross-Attention Using Graph-Based and Vision Backbone-Extracted Features

TL;DR

This paper introduces Tighnari, a multi-modal plant species prediction model that integrates hierarchical cross-attention and graph-based feature correction, utilizing diverse environmental data and advanced neural networks for improved biodiversity analysis.

Contribution

The work presents a novel hierarchical cross-attention mechanism and a graph-based feature correction method for multi-modal plant species prediction, leveraging multiple data sources and backbone networks.

Findings

Enhanced prediction accuracy demonstrated through ablation studies.

Effective fusion of temporal and image features via hierarchical cross-attention.

Robustness of the model validated with 10-fold cross-fusion experiments.

Abstract

Predicting plant species composition in specific spatiotemporal contexts plays an important role in biodiversity management and conservation, as well as in improving species identification tools. Our work utilizes 88,987 plant survey records conducted in specific spatiotemporal contexts across Europe. We also use the corresponding satellite images, time series data, climate time series, and other rasterized environmental data such as land cover, human footprint, bioclimatic, and soil variables as training data to train the model to predict the outcomes of 4,716 plant surveys. We propose a feature construction and result correction method based on the graph structure. Through comparative experiments, we select the best-performing backbone networks for feature extraction in both temporal and image modalities. In this process, we built a backbone network based on the Swin-Transformer Block for extracting temporal Cubes features. We then design a hierarchical cross-attention mechanism capable of robustly fusing features from multiple modalities. During training, we adopt a 10-fold cross-fusion method based on fine-tuning and use a Threshold Top-K method for post-processing. Ablation experiments demonstrate the improvements in model performance brought by our proposed solution pipeline.