Instrumentation & Measurement Magazine 24-3 - 83

low detection rate. This affects both our Deep SORT adaption and the IOU-Tracker. As mentioned above, the Music
Video dataset contains unlabeled faces. In our new benchmark all presented faces are annotated and have their IDs
assigned accordingly.
From the analysis above, although our multi-face detection
and tracking system performs fairly well in general, it still produces tracking errors such as IDS.

[9]	 J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, " You only look
once: unified, real-time object detection, " in Proc. IEEE Conf. Comp
Vision Pattern Recognition, pp. 779-788, 2016.
[10]	J. Huang, V. Rathod, C., Sun et al., " Speed/accuracy trade-offs for
modern convolutional object detectors, " in Proc. IEEE Conf. Comp
Vision Pattern Recognition, pp. 7310-7311, 2017.
[11]	W. Liu, D. Anguelov, D. Erhan, C. Szegedy, S. Reed, C. Y. Fu, and
A. C. Berg, " Ssd: single shot multibox detector, " in Proc. Eur. Conf.
Comp. Vision, pp. 21-37, 2016.

Conclusion

[12]	J. Dai, Y. Li, K. He, and J. Sun, " R-fcn: object detection via region-

Our work demonstrates that tracking-by-detection has
reached a performance level suitable for measurement tasks
when standard blocks for detection and data association are
adapted. We present an online multi-face tracking approach
based on the Deep SORT matching cascade. We review and
train MTCNN and two common object detectors with three
different feature extractors for face detection. We adapt Deep
SORT for tracking faces and explore different feature classifier loss functions. We contribute a benchmark dataset, the
House of Commons, for multi-face tracking in council and
parliament settings. While two state-of-the-art offline trackers produced better results in our evaluation, our approach is
online and real-time, and also performs similarly to the stateof-the-art high speed IOU-Tracker. Our tracker built with
standard blocks achieved fewer identity switches. It could also
be used for tracking in other measurement tasks, and it stands
to improve further as feature classifiers and face detectors improve further.

based fully convolutional networks, " Advances Neural Inf. Process.
Syst., pp. 379-387, 2016.
[13]	S. Shirmohammadi and A. Ferrero, " Camera as the instrument:
the rising trend of vision based measurement, " IEEE Instrum.
Meas. Mag., vol. 17, pp. 41-47, 2014.
[14]	L. De Vito, O. Postolache, and S. Rapuano, " Measurements
and sensors for motion tracking in motor rehabilitation, " IEEE
Instrum. Meas. Mag., vol. 17, pp. 30-38, 2014.
[15]	A. Mammeri, T. Zuo, and A. Boukerche, " Extending the detection
range of vision-based vehicular instrumentation, " IEEE Trans.
Instrum. Meas., vol. 65, pp. 856-873, 2016.
[16]	M. D. Cordea, D. C. Petriu, E. M. Petriu, N. D. Georganas, and T.
E. Whalen, " 3-D head pose recovery for interactive virtual reality
avatars, " IEEE Trans. Instrum. Meas., vol. 51, pp. 640-644, 2002.
[17]	G. Guo and X. Wang, " Kinship measurement on salient facial
features, " IEEE Trans. Instrum. Meas., vol. 61, pp. 2322-2325, 2012.
[18]	D. Lo, R. A. Goubran, R. M. Dansereau, G. Thompson, and
D. Schulz, " Robust joint audio-video localization in video
conferencing using reliability information, " IEEE Trans. Instrum.

References

Meas., vol. 53, pp. 1132-1139, 2004.

[1]	 S. Zhang, Y. Gong, J. B. Huang, J. Lim, J. Wang, N. Ahuja,

[19]	M. D. Cordea, E. M. Petriu, and D. C. Petriu, " Three-dimensional

and M. H. Yang, " Tracking persons-of-interest via adaptive

head tracking and facial expression recovery using an

discriminative features, " in Proc. Eur. Conf. Comp. Vision, pp. 415-

anthropometric muscle-based active appearance model IEEE

433, 2016.

Trans. Instrum. Meas., vol. 57, pp. 1578-1588, 2008.

[2]	 C. C. Lin and Y. Hung, " A prior-less method for multi-face

[20]	F. Deng, J. Dong, X. Wang, Y. Fang, Y. Liu, Z. Yu, J. Liu, and

tracking in unconstrained videos, " in Proc. IEEE Conf. Comp Vision

F. Chen, " Design and implementation of a noncontact sleep

Pattern Recognition, pp. 538-547, 2018.

monitoring system using infrared cameras and motion sensor, "

[3]	 N. Wojke, A. Bewley, and D. Paulus, " Simple online and realtime
tracking with a deep association metric, " in Proc. IEEE Int. Conf.
Image Process., pp. 3645-3649, 2017.

IEEE Trans. Instrum. Meas., vol. 67, pp. 1555-1563, 2018.
[21]	M. Omidyeganeh, S. Shirmohammadi, S. Abtahi et al., " Yawning
detection using embedded smart cameras, " IEEE Trans. Instrum.

[4]	 N. Wojke and A. Bewley, " Deep cosine metric learning for person
re-identification, " in Proc. IEEE Winter Conf. Applications of Comp.
Vision, pp. 748-756, 2018.

Meas., vol. 65, pp. 570-582, 2016.
[22]	P. Viola and M. Jones, " Robust real-time face detection, " in Proc.
IEEE Int. Conf. Comp. Vision, vol. 2, 2001.

[5]	 W. Liu, Y. Wen, Z. Yu, M. Li, B. Raj, and L. Song, " Sphereface: deep

[23]	X. Liu, X. Yang, D. Wang, and A. Wong, " Detecting pulse rates

hypersphere embedding for face recognition, " in Proc. IEEE Conf.

from facial videos recorded in unstable lighting conditions: an

Comp Vision Pattern Recognition, pp. 212-220, 2017.

adaptive spatio-temporal homomorphic filtering algorithm, "

[6]	 S. Engelberg and B. Milgrom, " Tracking using state estimation: a
brief introduction, " IEEE Instrum. Meas. Mag., vol. 22, pp. 36-42,
2019.

IEEE Trans. Instrum. Meas., Early Access, vol. 1-1, 2020.
[24]	T. Y. Lin, M. Maire, S. Belongie, J. Hays, P. Perona, D. Ramanan, P.
Dollar, and C. L. Zitnick, " Microsoft COCO: common objects in

[7]	 P. D. Z. Varcheie and G. Bilodeau, " Adaptive fuzzy particle filter
tracker for a PTZ camera in an IP surveillance system, " IEEE
Trans. Instrum. Meas., vol. 60, pp. 354-371, 2011.
[8]	 A. Antonucci, V. Magnago, L. Palopoli, and D. Fontanelli,
" Performance assessment of a people tracker for social robots, "

context, " in Proc. Eur. Conf. Comp. Vision, pp. 740-755, 2014.
[25]	C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, and Z. Wojna,
" Rethinking the inception architecture for computer vision, " in Proc.
IEEE Conf. Comp Vision Pattern Recognition, pp. 2818-2826, 2016.
[26]	K. He, X. Zhang, S. Ren, and J. Sun, " Deep residual learning

in Proc. 2019 IEEE Int. Instrum. Meas. Technol. Conf. (I2MTC), pp.

for image recognition, " in Proc. IEEE Conf. Comp Vision Pattern

1-6, 2019.

Recognition, pp. 770-778, 2016.

May 2021	

IEEE Instrumentation & Measurement Magazine	83
Instrumentation & Measurement Magazine 24-3

Table of Contents for the Digital Edition of Instrumentation & Measurement Magazine 24-3
