Mining visitors in Video Surveillance system

Momin, B. F.; Jere, Y.R.

doi:10.1109/iciiecs.2015.7193040

Search citation statements

Order By: Relevance

Paper Sections

Select...

Related Work1

Citation Types

Supporting

Mentioning

Contrasting

Year Published

2023

Publication Types

Select...

Other1

Relationship

Self Cite0

Independent1

Authors

Journals

Cited by 1 publication

(1 citation statement)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Early works for face selection utilize the Viola-Jones [7] model to detect faces in videos and, afterward, analyze each face according to quality metrics such as pose, brightness, resolution and detection confidence. These methods [4], [8], [9] demonstrate that choosing faces based on quality metrics increases recognition accuracy and super-resolution quality [10].…”

Section: Related Workmentioning

confidence: 96%

Modular Multi-Face Tracking Geared Toward Face Recognition in Surveillance Videos

Nascimento,

J. Neto,

Silva

2023

Anais Estendidos Da XXXVI Conference on Graphics, Patterns and Images (SIBRAPI Estendido 2023)

View full text Add to dashboard Cite

Face recognition has achieved great accuracy when used in controlled conditions, however, these results aren’t usually carried over to video surveillance scenarios. To facilitate the use of face recognition for video surveillance, face selection can be employed as an intermediate step. This dissertation presents a study of face selection where we rework a multi-face tracking pipeline and with few changes manage to increase tracking and reconnection capabilities. Through experimentation with different face detection models, random parameter search and a simpler face quality measure, we achieved an increase of 10.1% in Multiple Object Tracking Precision (MOTP) and 9% more in the IDF1 metric. All experiments were conducted on a public multi-face tracking dataset, which we also expanded through manual video annotations.

show abstract