Two-stage color fundus image registration via Keypoint Refinement and Confidence-Guided Estimation

Computerized Medical Imaging and Graphics, Journal Year: 2025, Volume and Issue: 123, P. 102554 - 102554

Published: April 19, 2025

Language: Английский

---

Fractal-Based Architectures with Skip Connections and Attention Mechanism for Improved Segmentation of MS Lesions in Cervical Spinal Cord

Diagnostics, Journal Year: 2025, Volume and Issue: 15(8), P. 1041 - 1041

Published: April 19, 2025

Background/Objectives: Multiple sclerosis (MS) is an autoimmune disease that damages the myelin sheath of central nervous system, which includes brain and spinal cord. Although MS lesions in are more frequently investigated, cervical cord (CSC) can be much specific for diagnosis disease. Furthermore, as lesion burden CSC directly related to progression, presence may help differentiate from other neurological diseases. Methods: In this study, two novel deep learning models based on fractal architectures proposed automatic detection segmentation by improving convolutional connection structures used layers U-Net architecture. our previous we introduced FractalSpiNet architecture incorporating block into framework develop a deeper network segmenting CPC. improve smaller finer details images, attention mechanism integrated architecture, resulting Att-FractalSpiNet model. addition, second hybrid model, incorporated skip structure development Con-FractalU-Net Results: Experimental studies were conducted using U-Net, FractalSpiNet, Con-FractalU-Net, detect region within its boundaries. region, achieved highest Dice Similarity Coefficient (DSC) score 98.89%. Similarly, detecting model again best performance with DSC 91.48%. Conclusions: For lesions, fractal-based higher scores than baseline particularly small complex structures.

Language: Английский

---

Brain tumor detection empowered with ensemble deep learning approaches from MRI scan images

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: April 29, 2025

Language: Английский

---

Hybrid Shunted Transformer embedding UNet for remote sensing image semantic segmentation

Neural Computing and Applications, Journal Year: 2024, Volume and Issue: 36(25), P. 15705 - 15720

Published: May 18, 2024

Language: Английский

---

Residual Echo State Networks: Residual recurrent neural networks with stable dynamics and fast learning

Neurocomputing, Journal Year: 2024, Volume and Issue: 597, P. 127966 - 127966

Published: May 31, 2024

Language: Английский

---