Pattern Matching Based on Scan Metrics for People Search and Face Detection
International Journal of Computer Studies and Application · 17.05.2022 · 4 (2) · 68
DOI: 10.66419/bgzt7973 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Computer Studies and Application, online erschienen am 16.05.2022. Verfasst von Fatima Isiaka und 1 weiteren Autor:innen. Die Arbeit nennt 1 Literaturangaben.
Abstract (Originalsprache)
The process uses a multi-tasking convolutional neural network (MTCNN) that trains the database model to identify and detect the best match to the input image. Each feature of the face is characterised by the eyes, nose and mouth detection in an acyclic pattern, each circle captures and intersects at the region of the nose which is the centroid of the face and this captures the entire face. The search pattern is based on a random search until the best fit is detected. Error in tracking (tracking operating characteristics (TOC)) was also computed for both face detection and emotion recognition as the two basic tasks. The error indicates that there is little margin for face identification compared to the emotion detected on the face, and further improvement would be to use to the model to determine and compare performance in the search and tracking processes.
Zitieren
APA
Isiaka, F., & Adamu, Z. (2022). Pattern Matching Based on Scan Metrics for People Search and Face Detection. International Journal of Computer Studies and Application, 4(2), 68. https://doi.org/10.66419/bgzt7973
Vancouver
Isiaka F, Adamu Z. Pattern Matching Based on Scan Metrics for People Search and Face Detection. International Journal of Computer Studies and Application. 2022;4(2):68. doi:10.66419/bgzt7973
BibTeX
@article{Isiaka2022,
author = {Isiaka, Fatima and Adamu, Zainab},
title = {Pattern Matching Based on Scan Metrics for People Search and Face Detection},
journal = {International Journal of Computer Studies and Application},
year = {2022},
volume = {4},
number = {2},
pages = {68},
doi = {10.66419/bgzt7973}
}
EUVIMED