We address the challenging problem of recognizing the camera wearer's actions from videos captured by an egocentric camera. Egocentric videos encode a rich set of signals regarding the camera wearer, including head movement, hand pose and gaze information. We propose to utilize these mid-level egocentric cues for egocentric action recognition. We present a novel set of egocentric features and show how they can be combined with motion and object features. The result is a compact representation with superior performance. In addition, we provide the first systematic evaluation of motion, object and egocentric cues in egocentric action recognition. Our benchmark leads to several surprising findings. These findings uncover the best practices for egocentric actions, with a significant performance boost over all previous state-of-the-art methods on three publicly available datasets.
Download full-text PDF |
Source |
|---|---|
| http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4784702 | PMC |
| http://dx.doi.org/10.1109/CVPR.2015.7298625 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
Right middle occipital gyrus is associated with egocentric spatial orientation during body tilt: Evidence from a repetitive transcranial magnetic stimulation study.
Eur J Neurosci
January 2025
Faculty of Psychology, Otemon Gakuin University, Ibaraki, Osaka, Japan.
Accurate perception of the orientation of external objects relative to the body, known as egocentric spatial orientation, is fundamental to performing action. Previously, we found via behavioural and magnetic resonance imaging voxel-based morphometry studies that egocentric spatial orientation is strongly distorted when the whole body is tilted with respect to gravity, and that the magnitude of this perceptual distortion is correlated with the grey matter volume of the right middle occipital gyrus (rMOG). In the present study, we further validated the association between the neural processing in the rMOG and the perceptual distortion by transiently suppressing neural activity in this region using low-frequency repetitive transcranial magnetic stimulation (rTMS) and evaluating the consequent effect on perceptual distortion.
View Article and Find Full Text PDFThe perspective during gestural executions alters hemispherical specialization.
Behav Brain Res
March 2025
Department of Neurology, Psychosomatic Medicine and Psychiatry, German Sport University (GSU) Cologne, Am Sportpark Muengersdorf 6, Cologne 50933, Germany.
Introduction: The left hemisphere may be particularly specialized for gestures from an egocentric movement perspective, i.e., when executing tool-use pantomime (TUP) gestures.
View Article and Find Full Text PDFBuilding egocentric models of local space from retinal input.
Curr Biol
December 2024
Behavioral and Systems Neuroscience, Department of Psychology, Rutgers University, New Brunswick, NJ 08854, USA. Electronic address:
Determining the location of objects relative to ourselves is essential for interacting with the world. Neural activity in the retina is used to form a vision-independent model of the local spatial environment relative to the body. For example, when an animal navigates through a forest, it rapidly shifts its gaze to identify the position of important objects, such as a tree obstructing its path.
View Article and Find Full Text PDFPersonalized Video-Based Hand Taxonomy Using Egocentric Video in the Wild.
IEEE J Biomed Health Inform
November 2024
Objective: Hand function is central to inter- actions with our environment. Developing a comprehen- sive model of hand grasps in naturalistic environments is crucial across various disciplines, including robotics, ergonomics, and rehabilitation. Creating such a taxonomy poses challenges due to the significant variation in grasp- ing strategies that individuals may employ.
View Article and Find Full Text PDFAltercentrism in perspective-taking: The role of humanisation in embodying the agent's point of view.
Q J Exp Psychol (Hove)
November 2024
Department of Psychology, University of Campania "Luigi Vanvitelli," Caserta, Italy.
Article Synopsis
- The study explored how the human-likeness and attractiveness of agents (human vs. robotic) affect participants' ability to adopt another's perspective in Visual Perspective-Taking (VPT) tasks.
- Participants were asked to judge the location of a target object based on different scenarios involving either human or robotic actors demonstrating various actions and cues, like gazing and stillness.
- Results showed that individuals could be grouped into two distinct styles of perspective-taking, influenced more by the nature of the agent than by attractiveness or social cues, highlighting the importance of understanding these factors for improving human-robot interactions.
Want AI Summaries of new PubMed Abstracts delivered to your In-box?
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!