Flame Image Classification Based on Deep Learning and Three-Way Decision-Making

Forests, Journal Year: 2025, Volume and Issue: 16(3), P. 544 - 544

Published: March 19, 2025

The classification and recognition of flame images play an important role in avoiding forest fires. Deep learning technology has shown good performance image tasks. In order to further improve the accuracy classification, this paper combines deep with idea three-way decision-making. First, a ResNet34 network is used for initial classification. probability value calculated by SoftMax function as decision evaluation criterion Using decision-making, divided into positive domain, negative boundary domain based on indicators. Furthermore, we perform secondary domains. DualArchClassNet structure was constructed extract new features combine them integrated are optimized reclassify uncertain domains overall accuracy. experimental results show that proposed method improves compared using single network.

Language: Английский

Cost-effective and accuracy-oriented ℓ1-norm support vector machine for enhanced feature selection

Measurement, Journal Year: 2025, Volume and Issue: unknown, P. 117506 - 117506

Published: April 1, 2025

Language: Английский