Tracking and Recognition of Multiple Faces at Distances

Liu, Rong; Gao, Xiufeng; Chu, Rufeng; Zhu, Xiangxin; Li, Stan Z.

doi:10.1007/978-3-540-74549-5_54

Cited by 8 publications

(2 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Generally speaking, single face tracking [18], [9], [15], [17], [20] does not require recognition, multiple faces tracking [14] demands accurate many-to-many recognition and video-based face recognition [10], [11] integrates tracking for better recognition. Differently, what the person-specific face tracking needs is one-to-many recognition.…”

Section: On-line Face Recognizermentioning

confidence: 99%

Person-specific face tracking with online recognition

Cai

Wen

Cao

et al. 2013

2013 10th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition (FG)

Self Cite

View full text Add to dashboard Cite

Abstract-Person-specific face tracking is a challenging task for the trackers which only focus on the appearance of the target face, because distraction always happens and the identity is difficult to maintain. In this paper, we design a framework combining an off-line detector, an on-line tracker and an online recognizer to complete the tracking of person-specific face. Recognizer is the key component in our framework, because the most confident target face will be selected by the recognizer from the pool of detected and tracked faces. Since there is no much prior information about the identities available and the face poses change frequently in surveillance scenarios, accurate recognition is extremely difficult and an on-line formulation is required. In order to ensure the precision of identity recognition with different poses, we project the extracted features of faces to a latent space with the help of Canonical Correlation Analysis (CCA) technique, and then these projected features are incrementally trained using an on-line SVM (LASVM). Experimental results demonstrate that our person-specific face tracking outperforms several state-of-the-art face trackers.

show abstract

Section: On-line Face Recognizermentioning

confidence: 99%

Person-specific face tracking with online recognition

Cai

Wen

Cao

et al. 2013

2013 10th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition (FG)

Self Cite

View full text Add to dashboard Cite

show abstract

“…Given a set of pan and tilt values, the MVC controls the velocity of pan and tilt motion. While there have been a few previous studies on the static image processing part [16], no systematic study has been reported on the dynamic camera control part.…”

Section: Camera Control For Smooth Trackingmentioning

confidence: 99%

PTZ camera assisted face acquisition, tracking & recognition

Choi

Park

Jain

2010

2010 Fourth IEEE International Conference on Biometrics: Theory, Applications and Systems (BTAS)

View full text Add to dashboard Cite

Face recognition systems typically have a rather short operating distance with standoff (distance between the camera and the subject) limited to 1∼2 meters. When these systems are used to capture face images at a larger distance (5∼10 m), the resulting images contain only a small number of pixels on the face region, resulting in a degradation in face recognition performance. To address this problem, we propose a camera system consisting of one PTZ camera and two static cameras to acquire high resolution face images up to a distance of 10 meters. We propose a novel camera calibration method based on the coaxial configuration between the static and PTZ cameras. We also use a linear prediction model and camera control to mitigate delays in image processing and mechanical camera motion. The proposed system has a larger standoff in face image acquisition and effectiveness in face recognition test. Experimental results on video data collected at a distance ranging from 5 to 10 meters of 20 different subjects as probe and 10,020 subjects as gallery shows 96.4% rank-1 identification accuracy of the proposed method compared to 0.1% rank-1 accuracy of the conventional camera system using a state-ofthe-art matcher.

show abstract