Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Feb 3;22(3):1153.
doi: 10.3390/s22031153.

A Novel Integration of Face-Recognition Algorithms with a Soft Voting Scheme for Efficiently Tracking Missing Person in Challenging Large-Gathering Scenarios

Adnan Nadeem  1 , Muhammad Ashraf  2 , Kashif Rizwan  2 , Nauman Qadeer  2 , Ali AlZahrani  1 , Amir Mehmood  3 , Qammer H Abbasi  4
Affiliations

A Novel Integration of Face-Recognition Algorithms with a Soft Voting Scheme for Efficiently Tracking Missing Person in Challenging Large-Gathering Scenarios

Adnan Nadeem et al. Sensors (Basel). .

Abstract

The probability of losing vulnerable companions, such as children or older ones, in large gatherings is high, and their tracking is challenging. We proposed a novel integration of face-recognition algorithms with a soft voting scheme, which was applied, on low-resolution cropped images of detected faces, in order to locate missing persons in a challenging large-crowd gathering. We considered the large-crowd gathering scenarios at Al Nabvi mosque Madinah. It is a highly uncontrolled environment with a low-resolution-images data set gathered from moving cameras. The proposed model first performs real-time face-detection from camera-captured images, and then it uses the missing person's profile face image and applies well-known face-recognition algorithms for personal identification, and their predictions are further combined to obtain more mature prediction. The presence of a missing person is determined by a small set of consecutive frames. The novelty of this work lies in using several recognition algorithms in parallel and combining their predictions by a unique soft-voting scheme, which in return not only provides a mature prediction with spatio-temporal values but also mitigates the false results of individual recognition algorithms. The experimental results of our model showed reasonably good accuracy of missing person's identification in an extremely challenging large-gathering scenario.

Keywords: integration of face-recognition algorithms; large-crowd gatherings; soft voting scheme; tracking missing persons.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Figure 1
Figure 1
Sampled frames of a video sequence showing the large crowd gathering in Al Nabvi mosque, Madinah (KSA).
Figure 2
Figure 2
Related work exploration in identifying missing persons in a huge crowd by face detection with low-resolution images [8,9,10,12,13,14,15,16,17,18].
Figure 3
Figure 3
Proposed cameras setups and spatial distribution of Al-Nabvi mosque, Madinah.
Figure 4
Figure 4
Proposed methodology workflow diagram.
Figure 5
Figure 5
Prediction maturity by soft voting.
Figure 6
Figure 6
Tracking in spatio-temporal context.
Figure 7
Figure 7
Detected face regions.
Figure 8
Figure 8
Extracted face regions: (a) cropped; (b) Enhanced & resized.
Figure 9
Figure 9
Personal identification by individual algorithms, where (ID, Score) pairs in every sub-caption indicate the face match and its score.
Figure 9
Figure 9
Personal identification by individual algorithms, where (ID, Score) pairs in every sub-caption indicate the face match and its score.
Figure 10
Figure 10
Prediction after maturity by soft-voting.
Figure 11
Figure 11
Tracking results of ID-142, where both rough and smooth presence tracks are shown. (The dashed lines represent manual and solid lines the system generated tracks).
Figure 11
Figure 11
Tracking results of ID-142, where both rough and smooth presence tracks are shown. (The dashed lines represent manual and solid lines the system generated tracks).
Figure 12
Figure 12
Performance evaluation for identifying and tracking ID-142.
Figure 13
Figure 13
Sampled personal images of size 50 × 50.
Figure 14
Figure 14
Sampled face images from the dataset, where (a) shows un-registered personals, (b) registered personals and (c) the registered missing personals.
Figure 15
Figure 15
Face-recognition comparative analysis.
Figure 16
Figure 16
Comparative analysis on accuracy.
Figure 17
Figure 17
Tracking results of 16 personals.
Figure 17
Figure 17
Tracking results of 16 personals.
Figure 18
Figure 18
Evaluation of personal tracking in temporal context. Recalls and accuracies were much lower than precisions. Proposed-approach performance was better than individual algorithms.
Figure 18
Figure 18
Evaluation of personal tracking in temporal context. Recalls and accuracies were much lower than precisions. Proposed-approach performance was better than individual algorithms.
Figure 19
Figure 19
Evaluation summary of temporal tracking for all registered personals.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Huang G.B., Ramesh M., Berg T., Learned-Miller E. Labeled Faces in the Wild: A Database for Studying Face Recognition in Unconstrained Environments. University of Massachusetts; Amherst, MA, USA: 2007.
    1. Kumar N., Berg A.C., Belhumeur P.N., Nayar S.K. Attribute and simile classifiers for face verification; Proceedings of the 2009 IEEE 12th International Conference on Computer Vision. IEEE; Kyoto, Japan. 29 September–2 October 2009; pp. 365–372.
    1. Pasztor E., Carmichael O.T. Learning low-level vision. Int. J. Comput. Vis. 2000;40:25–47.
    1. Liu Z., Luo P., Wang X., Tang X. Large-scale celebfaces attributes (celeba) dataset. Retrieved August. 2018;15:11.
    1. Wolf L., Hassner T., Maoz I. Face recognition in unconstrained videos with matched background similarity; Proceedings of the CVPR 2011; Colorado Springs, CO, USA. 20–25 June 2011; pp. 529–534.

LinkOut - more resources