Towards a practical PTZ face detection and tracking system

Cai, Yaqi; Medioni, Gérard; Dinh, Thang Ba

doi:10.1109/wacv.2013.6474996

Cited by 16 publications

(11 citation statements)

References 20 publications

(32 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Scheduling of a single PTZ camera was considered in [13,2,11,12]. Scenarios of joining/splitting targets were considered in [13,2]. The goal of [13] was to minimize the slew time of an aerial camera tracking cars.…”

Section: Previous Workmentioning

confidence: 99%

“…High-resolution images were used to remove incorrect prediction hypothesis (stored as a tree). The greedy policy in [2] aimed to maximize the number of captured faces, considering the predicted time of each target to exit the scene and its movement angle w.r.t. the camera.…”

Section: Previous Workmentioning

confidence: 99%

“…A prerequisite for choosing to zoom in on an unlabeled target is that the acquisition of its face is expected to be successful. A Boolean value that indicates the expected success, E s (v), can be computed by the tracker in a similar manner to previous studies (e.g., [2]). For example, the motion direction can be used for predicting occlusions and time to exit, and whether the face will be visible to the camera.…”

Section: Schedulingmentioning

confidence: 99%

“…According to the proposed scheduler, the system selects a person to zoom in on using a camera control algorithm. The control algorithm chooses the FOV that makes it possible to zoom in on selected target (e.g., [2]). In the zoom-in mode, a face image is acquired and a face-to-face and a face-to-person matchings are computed.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

The Multi-strand Graph for a PTZ Tracker

et al. 2017

Self Cite

View full text Add to dashboard Cite

High-resolution images can be used to resolve matching ambiguities between trajectory fragments (tracklets), which is one of the main challenges in multiple target tracking. A PTZ camera, which can pan, tilt and zoom, is a powerful and efficient tool that offers both close-up views and wide area coverage on demand. The wide-area view makes it possible to track many targets while the close-up view allows individuals to be identified from high-resolution images of their faces. A central component of a PTZ tracking system is a scheduling algorithm that determines which target to zoom in on.In this paper we study this scheduling problem from a theoretical perspective, where the high resolution images are also used for tracklet matching. We propose a novel data structure, the Multi-Strand Tracking Graph (MSG), which represents the set of tracklets computed by a tracker and the possible associations between them. The MSG allows efficient scheduling as well as resolving -directly or by elimination -matching ambiguities between tracklets. The main feature of the MSG is the auxiliary data saved in each vertex, which allows efficient computation while avoiding time-consuming graph traversal. Synthetic data simulations are used to evaluate our scheduling algorithm and to demonstrate its superiority over a naïve one.

show abstract

Section: Previous Workmentioning

confidence: 99%

Section: Previous Workmentioning

confidence: 99%

Section: Schedulingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

The Multi-strand Graph for a PTZ Tracker

et al. 2017

Self Cite

View full text Add to dashboard Cite

show abstract

“…To the best of our knowledge, few efforts have been done on this topic [1]. The second contribution of the paper is an online multiple target tracking algorithm.…”

Section: Introductionmentioning

confidence: 99%

Demo: Persistent people tracking and face capture using a PTZ camera

Cai

Medioni

2013

2013 Seventh International Conference on Distributed Smart Cameras (ICDSC)

Self Cite

View full text Add to dashboard Cite

In forensic applications, it is very helpful to capture the face images of the tracked people. However, regular CCTV cameras only capture a small number of pixels on the face region. A promising solution to this problem is to use a network of PTZ (Pan-Tilt-Zoom) cameras since the zoom capacities of PTZ cameras offer the option of a close view on demand. In this paper, we address the problem of persistent people tracking and automatic face capture using a single PTZ camera. The detected faces are associated with the corresponding people and trajectories. The time-critical and dynamic nature of our problem complicates our task. Different from previous work which use a mixture of wide angle cameras and PTZ cameras, we explore the limits to what can be expected from a single PTZ camera. The system first detects and tracks pedestrians in zoomed-out mode, then selects, using a scheduler, a person to zoom in. After zoom in, we come back to wide area mode, and solve the person-to-person, face-to-person and face-to-face data association problems. Extensive experiments in challenging outdoor uncontrolled conditions demonstrate the effectiveness of the proposed system.

show abstract