Advanced retinal disease detection from OCT images using a hybrid squeeze and excitation enhanced model

PLoS ONE, Journal Year: 2025, Volume and Issue: 20(2), P. e0318657 - e0318657

Published: Feb. 7, 2025

Retinal problems are critical because they can cause severe vision loss if not treated. Traditional methods for diagnosing retinal disorders often rely heavily on manual interpretation of optical coherence tomography (OCT) images, which be time-consuming and dependent the expertise ophthalmologists. This leads to challenges in early diagnosis, especially as diseases like diabetic macular edema (DME), Drusen, Choroidal neovascularization (CNV) become more prevalent. OCT helps ophthalmologists diagnose patients accurately by allowing detection. paper offers a hybrid SE (Squeeze-and-Excitation)-Enhanced Hybrid Model detecting from including DME, CNV, using artificial intelligence deep learning. The model integrates blocks with EfficientNetB0 Xception architectures, provide high success image classification tasks. achieves accuracy fewer parameters through scaling strategies, while powerful feature extraction separable convolutions. combination these architectures enhances both efficiency performance model, enabling accurate detection images. Additionally, increase representational ability network adaptively recalibrating per-channel responses. combined features processed via fully connected layers categorized Softmax algorithm. methodology was tested UCSD Duke's datasets produced excellent results. proposed SE-Improved outperformed current best-known approaches, rates 99.58% dataset 99.18% Duke dataset. These findings emphasize model's effectively images indicate substantial promise development computer-aided diagnostic tools field ophthalmology.

Language: Английский

Integrating Deep Learning and MRQy: A Comprehensive Framework for Early Detection and Quality Control of Brain Tumors in MRI Images using Python

International Journal of Computational and Experimental Science and Engineering, Journal Year: 2025, Volume and Issue: 11(2)

Published: April 15, 2025

The early detection of brain tumors is crucial for timely medical intervention and improved patient survival rates. Magnetic Resonance Imaging (MRI) the gold standard tumor diagnosis due to its superior soft-tissue contrast non-invasive nature. However, variations in MRI quality, including noise, artifacts, scanner inconsistencies, can impact diagnostic accuracy. This study aims de-velop a Python-based deep-learning model scans while integrating an automated quality control system using MRQy. MRQy, open-source tool, facilitates assessment by evaluating signal-to-noise ratios (SNR), contrast-to-noise (CNR), motion-related artifacts. deep learning will be trained on meticulously curated dataset, ensur-ing high-quality artifact-free images. By combining MRQy’s capabilities with techniques, expected en-hance accuracy reduce false-positive false-negative Furthermore, this research underscores significance standardized imaging protocols minimize variability across scanners institutions, ensuring repro-ducibility clinical AI applications. proposed approach leverages modern convolutional neural networks (CNNs) transfer incorpo-rating pre-trained architectures such as Res Net Efficient enhance fea-ture extraction. MRQy-based AI-driven classification, optimize MRI-based diagnostics, human error, improve outcomes. findings contribute ad-vancement AI-powered highlight importance

Language: Английский

PHOTODIAGNOSIS WITH DEEP LEARNING: A GAN AND AUTOENCODER-BASED APPROACH FOR DIABETIC RETINOPATHY DETECTION

Photodiagnosis and Photodynamic Therapy, Journal Year: 2025, Volume and Issue: unknown, P. 104552 - 104552

Published: March 1, 2025

Diabetic retinopathy (DR) is a leading cause of visual impairment and blindness worldwide, necessitating early detection accurate diagnosis. This study proposes novel framework integrating Generative Adversarial Networks (GANs) for data augmentation, denoising autoencoders noise reduction, transfer learning with EfficientNetB0 to enhance the performance DR classification models. GANs were employed generate high-quality synthetic retinal images, effectively addressing class imbalance enriching training dataset. Denoising further improved image quality by reducing eliminating common artifacts such as speckle noise, motion blur, illumination inconsistencies, providing clean consistent inputs model. was fine-tuned on augmented denoised The achieved exceptional metrics, including 99.00% accuracy, recall, specificity, surpassing state-of-the-art methods. custom-curated OCT dataset featuring high-resolution clinically relevant challenges limited annotated noisy inputs. Unlike existing studies, our work uniquely integrates GANs, autoencoders, EfficientNetB0, demonstrating robustness, scalability, clinical potential proposed framework. Future directions include interpretability tools adoption exploring additional imaging modalities improve generalizability. highlights transformative deep in critical diabetic

Language: Английский