Aerial Imagery Redefined: Next-Generation Approach to Object Classification

Information, Journal Year: 2025, Volume and Issue: 16(2), P. 134 - 134

Published: Feb. 11, 2025

Identifying and classifying objects in aerial images are two significant complex issues computer vision. The fine-grained classification of overhead has become widespread various real-world applications, due to recent advancements high-resolution satellite airborne imaging systems. task is challenging, particularly low-resource cases, the minor differences between classes within each class caused by nature. We introduce Classification Objects for Fine-Grained Analysis (COFGA), a recently developed dataset accurately categorizing images. COFGA comprises 2104 14,256 annotated across 37 distinct labels. This offers superior spatial information compared other publicly available datasets. MAFAT Challenge that utilizes improve methods. baseline model achieved mAP 0.6. cost was 60, whereas most score 0.6271 utilizing state-of-the-art ensemble techniques specific preprocessing techniques. offer solutions address difficulties analyzing images, when imbalanced data scarce. findings provide valuable insights into detailed categorization have practical applications urban planning, environmental assessment, agricultural management. discuss constraints potential future endeavors, specifically emphasizing integrate supplementary modalities contextual imagery analysis.

Language: Английский

Advanced Phishing Website Detection with SMOTETomek-XGB: Addressing Class Imbalance for Optimal Results

Procedia Computer Science, Journal Year: 2025, Volume and Issue: 252, P. 289 - 295

Published: Jan. 1, 2025

Language: Английский