CerviLearnNet: Advancing cervical cancer diagnosis with reinforcement learning-enhanced convolutional networks

Heliyon, Journal Year: 2024, Volume and Issue: 10(9), P. e29913 - e29913

Published: April 24, 2024

Women tend to face many problems throughout their lives; cervical cancer is one of the most dangerous diseases that they can face, and it has negative consequences. Regular screening treatment precancerous lesions play a vital role in fight against cancer. It becoming increasingly common medical practice predict early stages serious illnesses, such as heart attacks, kidney failure, cancer, using machine learning-based techniques. To overcome these obstacles, we propose use auxiliary modules special residual block, record contextual interactions between object classes support reference strategy. Unlike latest state-of-the-art classification method, create new architecture called Reinforcement Learning Cancer Network, "RL-CancerNet", which diagnoses with incredible accuracy. We trained tested our method on two well-known publicly available datasets, SipaKMeD Herlev, assess enable comparisons earlier methods. Cervical images were labeled this dataset; therefore, had be marked manually. Our study shows that, compared previous approaches for assignment classifying an cellular change, proposed approach generates more reliable stable image derived from datasets vastly different sizes, indicating will effective other datasets.

Language: Английский

---

ISTFormer: lightweight transformer for enhanced super-resolution of coal rock images via iterative feature extraction

The Visual Computer, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 3, 2025

Language: Английский

---

High-precision and lightweight small-target detection algorithm for low-cost edge intelligence

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Oct. 9, 2024

The proliferation of edge devices driven by advancements in Internet Things (IoT) technology has intensified the challenge achieving high-precision small target detection, as it demands extensive computational resources. This amplifies conflict between need for precise detection and requirement cost-efficiency across numerous devices. To solve this problem, paper introduces an enhanced algorithm, MSGD-YOLO, built upon YOLOv8. Faster Implementation CSP Bottleneck with 2 convolutions (C2f) module is through integration Ghost dynamic convolution, resulting a more lightweight architecture while enhancing feature generation. Additionally, Spatial Pyramid Pooling Enhanced Local Attention Network (SPPELAN) replaces Fast (SPPF) to expand receptive field, optimizing multi-level aggregation improved performance. Furthermore, novel Multi-Scale Convolution (MSGConv) Generalized Feature (MSGPFN) are introduced enhance fusion integrate multi-scale information. Finally, four optimized convolutional heads employed capture features accurately improve precision. Evaluation on VisDrone2019 dataset shows that compared YOLOv8-n, MSGD-YOLO improves mAP@50 mAP@50-95 14.1% 11.2%, respectively. In addition, model not only achieves 16.1% reduction parameters but also attains processing speed 24.6 Frames Per Second (FPS) embedded devices, thereby fulfilling real-time requirements.

Language: Английский

---

Dual-Branch Super-Resolution (DBSR): Lightweight DBSR Network for Enhancing Remote Sensing Images

Published: Feb. 19, 2025

Recent advances in deep learning-based super-resolution (SR) techniques for remote sensing images (RSIs) have shown significant promise. However, these performance improvements often come at a high computational cost, which limits their practical application. To address this issue, paper proposes dual-branch SR model (DBSR) that enhances both and efficiency through primary auxiliary branches. The branch integrates the advantages of channel recalibration, separable swin transformer (SST), spatial refinement module to achieve fine-grained feature extraction. SST serves as core branch, employing hierarchical window attention calculations facilitate lightweight effective multiscale representation. Conversely, shallow features global information enhancement module, mitigates misleading effects directly upsampling on results. Comparative ablation experiments conducted four RSI datasets five benchmark demonstrate our DBSR method effectively balances number parameters with performance, showcasing its potential application processing.

Language: Английский

---

PixMed-Enhancer: An Efficient Approach for Medical Image Augmentation

Bioengineering, Journal Year: 2025, Volume and Issue: 12(3), P. 235 - 235

Published: Feb. 26, 2025

AI-powered medical imaging faces persistent challenges, such as limited datasets, class imbalances, and high computational costs. To overcome these barriers, we introduce PixMed-Enhancer, a novel conditional GAN that integrates the ghost module into its encoder—a pioneering approach achieves efficient feature extraction while significantly reducing complexity without compromising performance. Our method features hybrid loss function, uniquely combining binary cross-entropy (BCE) Structural Similarity Index Measure (SSIM), to ensure pixel-level precision enhancing perceptual realism. Additionally, use of input masks offers unparalleled control over generation tumor features, marking breakthrough in fine-grained dataset augmentation for segmentation diagnostic tasks. Rigorous testing on diverse datasets establishes PixMed-Enhancer state-of-the-art solution, excelling realism, structural fidelity, efficiency. robust foundation real-world clinical applications AI-driven imaging.

Language: Английский

---

Advancing white balance correction through deep feature statistics and feature distribution matching

Journal of Visual Communication and Image Representation, Journal Year: 2025, Volume and Issue: unknown, P. 104412 - 104412

Published: Feb. 1, 2025

Language: Английский

---

DiffOSR: Latitude-aware conditional diffusion probabilistic model for omnidirectional image super-resolution

Knowledge-Based Systems, Journal Year: 2025, Volume and Issue: 315, P. 113244 - 113244

Published: March 2, 2025

---

Humanlike-GAN: a two-stage asymmetric CycleGAN for underwater image enhancement

Signal Image and Video Processing, Journal Year: 2025, Volume and Issue: 19(5)

Published: March 10, 2025

Language: Английский

---

HFD-NET: a real-time deep learning-based visual algorithm with weak feature enhancement for hair follicle detection

Signal Image and Video Processing, Journal Year: 2025, Volume and Issue: 19(5)

Published: March 12, 2025

Language: Английский

---

PixelCraftSR: Efficient Super-Resolution with Multi-Agent Reinforcement for Edge Devices

Sensors, Journal Year: 2025, Volume and Issue: 25(7), P. 2242 - 2242

Published: April 2, 2025

Single-image super-resolution imaging methods are increasingly being employed owing to their immense applicability in numerous domains, such as medical imaging, display manufacturing, and digital zooming. Despite widespread usability, the existing learning-based (SR) computationally expensive inefficient for resource-constrained IoT devices. In this study, we propose a lightweight model based on multi-agent reinforcement-learning approach that employs multiple agents at pixel level construct images by following asynchronous actor-critic policy. The iteratively select predefined set of actions be executed within five time steps new image state, followed action maximizes cumulative reward. We thoroughly evaluate compare our proposed method with methods. Experimental results illustrate can outperform models both qualitative quantitative scores despite having significantly less computational complexity. practicability is confirmed further evaluating it platforms, including edge

Language: Английский

---