Search papers, labs, and topics across Lattice.
The paper introduces FEEL, a large-scale dataset of egocentric video paired with force measurements from custom piezoresistive gloves, capturing natural hand-object manipulation in kitchen environments. They demonstrate the utility of force data for physical action understanding by applying FEEL to contact understanding (temporal and pixel-level segmentation) and action representation learning via force prediction as a self-supervised pretraining objective. Results show state-of-the-art temporal contact segmentation and improved transfer performance on action understanding tasks compared to training without force data.
Force sensing gloves unlock a new dimension of self-supervision for video models, boosting action understanding without manual labels.
We introduce FEEL (Force-Enhanced Egocentric Learning), the first large-scale dataset pairing force measurements gathered from custom piezoresistive gloves with egocentric video. Our gloves enable scalable data collection, and FEEL contains approximately 3 million force-synchronized frames of natural unscripted manipulation in kitchen environments, with 45% of frames involving hand-object contact. Because force is the underlying cause that drives physical interaction, it is a critical primitive for physical action understanding. We demonstrate the utility of force for physical action understanding through application of FEEL to two families of tasks: (1) contact understanding, where we jointly perform temporal contact segmentation and pixel-level contacted object segmentation; and, (2) action representation learning, where force prediction serves as a self-supervised pretraining objective for video backbones. We achieve state-of-the-art temporal contact segmentation results and competitive pixel-level segmentation results without any need for manual contacted object segmentation annotations. Furthermore we demonstrate that action representation learning with FEEL improves transfer performance on action understanding tasks without any manual labels over EPIC-Kitchens, SomethingSomething-V2, EgoExo4D and Meccano.
