Epilepsi Nöbet Tespiti için Zaman-Frekans Görüntüleme: Transformer Model ile Özellik Füzyonu

Mühendislik Bilimleri ve Araştırmaları Dergisi, Journal Year: 2025, Volume and Issue: 7(1), P. 93 - 102

Published: April 28, 2025

Epilepsi, nöbetler ve bu durumun yol açtığı geri dönüşümsüz beyin hasarı gibi ciddi riskler taşıyan yaygın bir nörolojik hastalıktır. Bu hastalığın doğru hızlı şekilde teşhis edilmesi büyük önem taşır. Geleneksel EEG sinyal analizi, manuel zaman alıcı olup insan hatalarına açıktır. sorunu çözmek için yapay zekâ yaklaşımlarının kullanımı, daha hassas tespit imkânı sunmaktadır. çalışmada, sinyalleri zaman-frekans dönüşüm yöntemleri kullanarak 2B görüntülere dönüştürülmüştür. Zaman-frekans ile üç adet görüntü kümesi elde edilmiştir. Ardından her transformer model eğitilmiştir tarafından özellik setleri oluşturulmuştur. Özellik füzyonu yöntemiyle farklı birleştirilmiş birleşik setler, makine öğrenmesi (destek vektör makineleri) sınıflandırılmıştır. çalışmada önerilen yaklaşım sayesinde %91.20 genel doğruluk oranı

---

Unsupervised Deep Learning for Flood Segmentation in UAV Imagery

Lecture notes in computer science, Journal Year: 2025, Volume and Issue: unknown, P. 408 - 423

Published: Jan. 1, 2025

Language: Английский

---

Small water body extraction in remote sensing with enhanced CNN architecture

Applied Soft Computing, Journal Year: 2024, Volume and Issue: unknown, P. 112544 - 112544

Published: Dec. 1, 2024

Language: Английский

---

A new hybrid mobile CNN approach for crosswalk recognition in autonomous vehicles

Multimedia Tools and Applications, Journal Year: 2024, Volume and Issue: 83(26), P. 67747 - 67762

Published: Jan. 27, 2024

Abstract While automobile transportation is increasing worldwide, it also negatively affects the safety of road users. Along with neglect traffic rules, pedestrians account for 22% all highway deaths. Millions suffer non-fatal injuries from these accidents. Most and deaths occur at crosswalks, where intersect. In this study, deep learning-based a new hybrid mobile CNN approaches are proposed to reduce by automatically recognizing crosswalks in autonomous vehicles. The first HMCNet approach, which model MobileNetv3 MNasNet models used together. This achieves approximately 2% more accuracy than peak performance lean models. Another approach FHMCNet increases success approach. LSVC feature selection method SVM classification addition HMCNet. increased 2%. Finally, offered 3% state-of-the-art methods literature.

Language: Английский

---

Unsupervised Color Based Flood Segmentation in UAV Imagery

Published: April 16, 2024

We propose a novel unsupervised semantic segmentation method for fast and accurate flood area detection utilizing color images acquired from Unmanned Aerial Vehicles (UAVs). To the best of our knowledge, this is first fully in captured by UAVs, without need pre-disaster images. The proposed framework addresses problem based on parameter-free calculated masks image analysis techniques. First, algorithm gradually excludes areas classified as non-flood over each component LAB colorspace, well an RGB vegetation index detected edges original image. Unsupervised techniques, such distance transform, are then applied, producing probability map location flooded areas. Finally, obtained applying hysteresis thresholding segmentation. tested compared with variations, other supervised methods two public datasets, consisting 953 total, yielding high-performance results, 87.4% 80.9% overall accuracy F1-Score, respectively. results computational efficiency show that it suitable board data execution decision-making during UAVs flight.

Language: Английский

---

A multi-level damage assessment model based on change detection technology in remote sensing images

Natural Hazards, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 26, 2024

Language: Английский

---

Unsupervised Color-Based Flood Segmentation in UAV Imagery

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(12), P. 2126 - 2126

Published: June 12, 2024

We propose a novel unsupervised semantic segmentation method for fast and accurate flood area detection utilizing color images acquired from unmanned aerial vehicles (UAVs). To the best of our knowledge, this is first fully in captured by UAVs, without need pre-disaster images. The proposed framework addresses problem based on parameter-free calculated masks image analysis techniques. First, algorithm gradually excludes areas classified as non-flood, over each component LAB colorspace, well using an RGB vegetation index detected edges original image. Unsupervised techniques, such distance transform, are then applied, producing probability map location flooded areas. Finally, obtained applying hysteresis thresholding segmentation. tested compared with variations other supervised methods two public datasets, consisting 953 total, yielding high-performance results, 87.4% 80.9% overall accuracy F1-score, respectively. results computational efficiency show that it suitable onboard data execution decision-making during UAV flights.

Language: Английский

---

A Multi-Level Damage Assessment Framework for Mixed-Hazard Buildings with Global Spatial Feature Guidance Module and Change Feature Attention in VHR Remote Sensing Images

IEEE Geoscience and Remote Sensing Letters, Journal Year: 2024, Volume and Issue: 21, P. 1 - 5

Published: Jan. 1, 2024

Language: Английский

---

Assessing the environmental impacts of flooding in Brazil using the flood area segmentation network deep learning model

Natural Hazards, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 6, 2024

Language: Английский

---