Multi-view Face Detection with One Classifier for Video Analytics Systems

Date

2014-08-24

Author

Kaya, Tunç Güven

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

121
views

0
downloads

In a video analytics for audience measurement system, dwell time, gaze, and opportunity-to-see statistics are required most of the time. To generate these statistics, more than one face detector is used in order to capture both profile and frontal faces. In this paper, we present a novel approach for face detection in video analytics. The assumption is that; the face occurrences are limited in such systems and one classifier is able to capture all of these occurrences. By using MB-LBP for feature extraction and Gentle Boost for statistical learning we trained a classifier which is able to detect both profiles and frontal faces with more than 80 % success rate. The proposed system also uses an efficient scanning algorithm and achieves 60 fps on a 720p video.

Subject Keywords

Multi-view face detection, Object detection, Audience measurement system, Multi-block lbp, Video analytics

URI

https://hdl.handle.net/11511/32354

DOI

https://doi.org/10.1007/978-3-319-12811-5_7

Collections

Graduate School of Informatics, Conference / Seminar

Suggestions

OpenMETU
Core

Rollout Algorithms for the Measurement-to-Track Association Problem Ozgen, Selim; Demirekler, Mübeccel (2016-05-19) The quality and precision of tracking manuevering targets under large clutter is highly dependent on both the data association and state estimation algorithms. In this study, measurement-to track association problem was discussed and the optimal association problem was shown to be a Markov Decision Process. The problem model considers the batch measurements in a time interval. The optimization problem has an heavy computational load, therefore the rollout algorithm is used to solve this problem. The approxi...
Event detection on social media using transaction based stream processing engine Çınar, Hüseyin Alper; Karagöz, Pınar; Department of Computer Engineering (2019) The aim of this study is detecting events on social media by improving current solutions in terms of accuracy and time performance. An event is something that occurs in a short duration of time in a certain place. In this thesis, the problem is modelled as a streaming transaction process. Three different event detection method is adapted to our solution. First one is the keyword-based event detection method that looks for bursty keywords in a period. The second one is the clustering-based event detection me...
Machine Learning Methods for Attack Detection in the Smart Grid Ozay, Mete; Esnaola, Inaki; Yarman Vural, Fatoş Tunay; Kulkarni, Sanjeev R.; Poor, H. Vincent (2016-08-01) Attack detection problems in the smart grid are posed as statistical learning problems for different attack scenarios in which the measurements are observed in batch or online settings. In this approach, machine learning algorithms are used to classify measurements as being either secure or attacked. An attack detection framework is provided to exploit any available prior knowledge about the system and surmount constraints arising from the sparse structure of the problem in the proposed approach. Well-known...
PL FSM: An Approach and a Tool for the Application of FSM in SPL Environments Eren, Onder; ÖZKAN, BARIŞ; Demirörs, Onur (2015-08-28) In order to develop cost-efficient software it is crucial to measure the accurate software size. However; measuring the software size has up to now been almost entirely a manual process and, as such, is both time-consuming and prone to human error which can end up with time and money loss. Automation of this process is a must for the software developing companies to improve the quality of project and budget planning. This paper introduces a mapping between COSMIC concept elements and UML conceptual elements...
Towards designing audio assistance for comprehending haptic graphs A multimodal perspective DENİZ, OZAN; MEHMETCAN, FAL; Acartürk, Cengiz (2013-07-31) Statistical graphs, such as line graphs are widely used in multimodal communication settings. Language accompanies graphs and humans produce gestures during the course of communication. For visually impaired people, haptic-audio interfaces provide perceptual access to graphical representations. The local and sequential character of haptic perception introduces limitations in haptic perception of hard-to-encode information, which can be resolved by providing audio assistance. In this article we first present...

Citation Formats

T. G. Kaya, “Multi-view Face Detection with One Classifier for Video Analytics Systems,” 2014, vol. 8811, Accessed: 00, 2020. [Online]. Available: https://hdl.handle.net/11511/32354.