Description
A method that randomly mask out all features coming from a specific sensor in multi-sensor models for Earth observation. Depending on the fusion strategy, the mask out can be done at the input, feature or decision level.
Papers Using This Method
Increasing the Robustness of Model Predictions to Missing Sensors in Earth Observation2024-07-22DifFUSER: Diffusion Model for Robust Multi-Sensor Fusion in 3D Object Detection and BEV Segmentation2024-04-06Incomplete Multimodal Learning for Remote Sensing Data Fusion2023-04-22Self-supervised Vision Transformers for Joint SAR-optical Representation Learning2022-04-11