Ego-surfing first person videos

  title={Ego-surfing first person videos},
  author={Ryo Yonetani and Kris Kitani and Yoichi Sato},
  journal={2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
We envision a future time when wearable cameras (e.g., small cameras in glasses or pinned on a shirt collar) are worn by the masses and record first-person point-of-view (POV) videos of everyday life. While these cameras can enable new assistive technologies and novel research challenges, they also raise serious privacy concerns. For example, first-person videos passively recorded by wearable cameras will necessarily include anyone who comes into the view of a camera - with or without consent… 

Figures and Tables from this paper

Egocentric Video Biometrics

It is shown that motion features in egocentric video provide biometric information, and the identity of the user can be reliably determined from a few seconds of video captured when the user is walking, which may include theft prevention of wearable cameras by locking the camera when not worn by its lawful owner.

Egocentric Meets Top-View

This study addresses two basic yet challenging questions about having a set of egocentric videos and a top-view surveillance video, and evaluates the capability of the proposed approaches in terms of jointly addressing the temporal alignment and assignment tasks.

An Egocentric Look at Video Photographer Identity

It is shown that camera motion, as can be computed from the egocentric video, provides unique identity information, and the photographer can be reliably recognized from a few seconds of video captured when walking.

Identifying First-Person Camera Wearers in Third-Person Videos

A new semi-Siamese Convolutional Neural Network architecture is proposed to address the challenge of establishing person-level correspondences across first-and third-person videos, which is challenging because the camera wearer is not visible from his/her own egocentric video, preventing the use of direct feature matching.

Anonymizing Egocentric Videos

A novel technique to anonymize egocentric videos is suggested, which create carefully crafted, but small, and imperceptible optical flow perturbations in an Egocentric video’s frames but are strong enough to dis-balance the gait recovery process.

Egocentric Meets Surveillance

This effort formalizes the problem in a way which handles and also estimates the unknown relative time-delays between the egocentric videos and the top-view video, and formulate the problem as a spectral graph matching instance, and jointly seek the optimal assignments and relativeTime-Delays of the videos.

Joint Person Segmentation and Identification in Synchronized First- and Third-person Videos

The proposed method performs significantly better than the state-of-the-art on both person segmentation and identification and is mutually beneficial, because better fine-grained segmentations allow for better matching across views, and using information from multiple views helps us perform more accurate segmentation.

EgoScanning: Quickly Scanning First-Person Videos with Egocentric Elastic Timelines

This work presents EgoScanning, a novel video fast-forwarding interface that helps users to find important events from lengthy first-person videos recorded with wearable cameras continuously. This

Integrating Egocentric Videos in Top-View Surveillance Videos: Joint Identification and Temporal Alignment

This paper proposes a unified framework to jointly solve all three problems of identification and temporal alignment of egocentric and top-view surveillance cameras, and evaluates the efficacy of the proposed approach on a publicly available dataset containing a variety of videos recorded in different scenarios.

EgoReID: Cross-view Self-Identification and Human Re-identification in Egocentric and Surveillance Videos

This work proposes a CRF-based method to identify the cameraman in the content of the top-view video, and also re-identify the people visible in the egocentric video, by matching them to the identities present in theTop view video.



Head Motion Signatures from Egocentric Videos

It is proposed to create a unique signature, based on pattern of head motion, that could identify that the subject is indeed the person seen in a video, and can identify the subject only at a particular time.

Head Pose Estimation in First-Person Camera Views

A new method for head pose real-time estimation in ego-vision scenarios that is a key step in the understanding of social interactions and uses and extends the Hough-Based Tracker to robustly detect head under changing aspect ratio, scale and orientation.

Person re-identification with a PTZ camera: An introductory study

An introductory study that paves the way for a new kind of person re-identification, by exploiting a single Pan-Tilt-Zoom (PTZ) camera, and shows that the proposed compound of two images overwhelms standard multi-shot descriptions, composed by many more pictures.

Gaze-enabled egocentric video summarization via constrained submodular maximization

This paper forms a summarization model which captures common-sense properties of a good summary, and shows that it can be solved as a submodular function maximization with partition matroid constraints, opening the door to a rich body of work from combinatorial optimization.

Wearable navigation system for the blind people in dynamic environments

  • Ya TianYong LiuJindong Tan
  • Computer Science
    2013 IEEE International Conference on Cyber Technology in Automation, Control and Intelligent Systems
  • 2013
A novel method based on feature points from a video sequence is proposed to not only estimate the camera motion itself but also the 3D motion of the moving object so as to infer the depth between the camera and theMoving object.

From Ego to Nos-Vision: Detecting Social Relationships in First-Person Views

A novel approach to detect groups in ego-vision scenarios by applying a correlation clustering algorithm that merges pairs of people into socially related groups using Structural SVMs.

Action Recognition with Improved Trajectories

  • Heng WangC. Schmid
  • Computer Science
    2013 IEEE International Conference on Computer Vision
  • 2013
Dense trajectories were shown to be an efficient video representation for action recognition and achieved state-of-the-art results on a variety of datasets are improved by taking into account camera motion to correct them.

Social interactions: A first-person perspective

Encouraging results on detection and recognition of social interactions in first-person videos captured from multiple days of experience in amusement parks are demonstrated.

PlaceAvoider: Steering First-Person Cameras away from Sensitive Spaces

This work introduces PlaceAvoider, a technique for owners of first-person cameras to ‘blacklist’ sensitive spaces (like bathrooms and bedrooms), and tests the technique on five realistic firstperson image datasets and shows it is robust to blurriness, motion, and occlusion.

Detecting bids for eye contact using a wearable camera

A learning-based method that couples a pose-dependent appearance model with a temporal Conditional Random Field (CRF) enables measuring gaze behavior in naturalistic social interactions and outperforms state-of-the-art approaches.