Affiliation:
1. Portuguese Air Force Academy Research Center, 2715-021 Pero Pinheiro, Portugal
2. Institute for Systems and Robotics, Instituto Superior Técnico, 1049-001 Lisboa, Portugal
Abstract
This work proposes an active learning (AL) methodology to create models for the segmentation of fire and smoke in video images. With this model, a model learns in an incremental manner over several AL rounds. Initially, the model is trained in a given subset of samples, and in each AL round, the model selects the most informative samples to be added to the training set in the next training session. Our approach is based on a decomposition of the task in an AL classification phase, followed by an attention-based segmentation phase with class activation mapping on the learned classifiers. The use of AL in classification and segmentation tasks resulted in a 2% improvement in accuracy and mean intersection over union. More importantly, we showed that the approach using AL achieved results similar to non-AL with fewer labeled data samples.
Subject
General Earth and Planetary Sciences
Reference34 articles.
1. Instituto da Conservação da Natureza e das Florestas (ICNF) (2021). Incêndios Rurais e área Ardida-Continente, ICNF.
2. Instituto da Conservação da Natureza e das Florestas (2016). Relatório anual de áreas ardidas e incêndios florestais em Portugal Continental, Technical Report.
3. See No Evil: Cognitive Challenges of Security Surveillance and Monitoring;Hodgetts;J. Appl. Res. Mem. Cogn.,2017
4. Lin, T.Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Dollár, P., and Zitnick, C.L. (2014, January 6–12). Microsoft coco: Common objects in context. Proceedings of the Computer Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland.
5. Santana, B., Cherif, E.K., Bernardino, A., and Ribeiro, R. (2022). Real-Time Georeferencing of Fire Front Aerial Images Using Iterative Ray-Tracing and the Bearings-Range Extended Kalman Filter. Sensors, 22.
Cited by
1 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献