
Neuroscience-inspired motion energy models outperform state-of-the-art deep learning optical flow models in zero-shot generalization to random dot stimuli for object segmentation, suggesting a closer alignment with human visual processing.


coremsg

motion-energy-processing-enables-human-like-zero-shot-generalization-to-random-dot-stimuli-for-object-segmentation


Motion Energy Processing Enables Human-like Zero-Shot Generalization to Random Dot Stimuli for Object Segmentation


title_rewrite


This paper introduces a novel unsupervised method for segmenting moving objects from aerial platforms using event cameras and self-supervised vision transformers, achieving state-of-the-art performance on multiple benchmarks.


unsupervised-event-based-motion-segmentation-for-aerial-surveillance-using-self-supervised-vision-transformers


Unsupervised Event-Based Motion Segmentation for Aerial Surveillance Using Self-Supervised Vision Transformers
