This review provides a comprehensive synthesis of efficient and lightweight deep learning architectures specifically tailored for the medical domain, and examines key model compression strategies and their efficacy in maintaining diagnostic performance while reducing hardware requirements.
Abstract
Deep learning has revolutionized medical image analysis, playing a vital role in modern clinical applications. However, the deployment of large-scale models in real-world clinical settings remains challenging due to high computational costs, latency constraints, and patient data privacy concerns associated with cloud-based processing. To address these bottlenecks, this review provides a comprehensive synthesis of efficient and lightweight deep learning architectures specifically tailored for the medical domain. We categorize the landscape of modern efficient models into three primary streams: Convolutional Neural Networks (CNNs) for local inductive bias, Lightweight Transformers for global context under constrained attention, and emerging Linear Complexity Models for scalable global context. Furthermore, we examine key model compression strategies (including pruning, quantization, knowledge distillation, and low-rank factorization) and evaluate their efficacy in maintaining diagnostic performance while reducing hardware requirements. By identifying current limitations and discussing the transition toward on-device intelligence, this review serves as a roadmap for researchers and practitioners aiming to bridge the gap between high-performance AI and resource-constrained clinical environments.
Accurate diagnosis from medical images is critical yet challenging in resource-constrained clinical settings, where models must balance high accuracy with computational efficiency for practical deployment. While deep learning has revolutionized medical image analysis, a significant gap persists: heavyweight models (e.g., Transformers, dense CNNs) achieve strong performance but are infeasible for edge devices, whereas existing lightweight architectures often lack the sophisticated context modeling needed for robust detection of subtle and heterogeneous pathologies. To address this challenges, we introduce the Context-Enhanced Network (CENet), an efficient convolutional architecture designed for medical image classification. CENet incorporates three key modules: (1) a Two-Level Feature Fusion Module (TLFM) that performs selective, gated fusion of only the most semantically complementary feature scales to prevent noise propagation; (2) a novel Triplet Excitation Attention (TEA), a parameter-efficient channel attention module for robust feature re-calibration; and (3) Lightweight Context-Enhanced Module (LightCEM) that enables iterative, self-guided feature refinement using global context. Extensive experiments across three medical imaging benchmarks, two brain tumor classification datasets (SARTAJ, Br35H) and dental radiography analysis demonstrate that CENet variants achieve state-of-the-art efficiency-accuracy trade-off. Notably, the CENet-small variant achieves weighted F1-score of 76.08% on the multi-class SARTAJ dataset while utilizing only 2.08 M parameters and 0.232 GFLOPs, outperforming larger architectures. Furthermore, Qualitative analyses using Grad-CAM further show that CENet consistently localizes clinically meaningful regions, enhancing interpretability. Meanwhile, CENet demonstrates real-time embedded performance on the NVIDIA Jetson Orin NX, achieving up to 75.17 FPS for CENet-small (and 62.21 FPS for CENet-large), which supports practical deployment under the strict latency and power constraints. While CENet achieves strong weighted performance across all benchmarks, classes with diffuse morphology (e.g., glioma) or subtle presentation (e.g., cavity) remain challenging and require further investigation. The code for this work will be made available at https://github.com/aminabenabid/CENet
Amina Benabid, Kangjie Cheng, Yun-Feng Liu et al.· Journal of King Saud Univers...· 0 citations
Empirical evidence from recent studies demonstrates that fine-tuning and network-based DTL strategies, including federated learning, consistently enhance diagnostic accuracy, robustness, and generalization across multiple medical imaging modalities, particularly in data-limited clinical scenarios.
M. A. S. Banu, A. Dhavapandiammal, K. Palanisamy· Current medical imaging· 0 citations
ABSTRACT
The adoption of whole-slide imaging is establishing a new paradigm in digital pathology. However, the translation of artificial intelligence (AI) from research to clinical practice faces significant hurdles, largely due to a misalignment between algorithmic advances and the practical demands of pathological diagnosis and prognosis. In this review, we propose a dual-perspective framework to systematically bridge this gap by linking core clinical tasks with cutting-edge deep learning methodologies. We present a comprehensive overview of the field from 2020 to 2025, analyzing how architectures such as convolutional neural networks, vision transformers, and graph neural networks are being adapted for diagnostic classification, tissue segmentation, and prognostic prediction. A key contribution is our novel algorithm-clinical task mapping framework, which offers practical guidance for selecting and designing AI solutions tailored to specific clinical goals. We also highlight emerging trends that minimize reliance on costly annotations-including weakly supervised and self-supervised learning-as well as advances in predicting immunohistochemistry results directly from hematoxylin and eosin-stained slides. Finally, we address critical challenges related to model interpretability, regulatory approval, and multicenter generalization, and outline a future pathway focused on developing integrated, trustworthy, and equitable AI systems that enhance, rather than replace, the expertise of pathologists.
Yun-qiu Gao, Teng Ma, Lisha Li et al.· Chinese Medical Journal· 0 citations
Major deep learning architectures, including CNNs, residual networks, UNet, attention-based models, Vision Transformers, and hybrid approaches, along with their clinical applications are summarized and emerging directions such as self-supervised learning, Explainable AI, federated learning, and lightweight models are highlighted as promising approaches for more reliable and accessible medical image analysis.
Lakshmi Sai Anusha Dadi, Pravallika Devi Kommana· International Journal for Re...· 0 citations
Oral cancer is a leading cause of mortality in low-to-middle-income countries, where a shortage of specialists delays diagnosis. While point-of-care screening via smartphones offers a scalable solution, developing robust AI for resource-constrained settings poses significant challenges, including class imbalance in training data, variable data quality, and computational constraints on edge devices. In this paper, we present the optimisation of lightweight deep learning models for smartphone-based oral cancer screening. Using a diverse, multi-centre retrospective dataset of approximately 30,000 images acquired over a decade, we systematically evaluate state-of-the-art convolutional, transformer, and hybrid architectures. Through rigorous pipeline ablation, we demonstrate that directly optimising hybrid architectures for the edge strictly outperforms computationally heavy paradigms, such as large models or knowledge distillation. Furthermore, interpretability analysis and simulated noise-stress tests revealed that the system anchors on clinical features and remains robust to unstructured sensor noise, despite vulnerabilities to impulse bit errors. In the held-out test set, our optimised MobileViTv2 models achieved an average sensitivity of 83.2 $\pm$ 1.5% and an average specificity of 86.0 $\pm$ 0.8%, with the best model exhibiting 87.4% sensitivity, 86.5% specificity, and a critical negative predictive value of 97.2% with reference to specialist labels. These results confirm that with targeted architectural selection and streamlined optimisation, interpretable and robust lightweight AI models exhibit high potential for edge deployment to enable automated triage in primary care settings.
Siddhant Bharadwaj, Aakash Shedsale, T. Subramanya et al.· 0 citations
Machine learning has become increasingly important in medical diagnosis, yet its effectiveness depends on access to large, reliable, and high-quality datasets. During epidemics and emerging diseases, such as COVID-19, acquiring sufficient real-world medical images rapidly is challenging. To address these issues, this study presents an Adaptive Conditional Generative Adversarial Network (Adaptive-CGAN) integrated with a cloud-based medical image processing framework. The proposed approach makes three main contributions. First, Adaptive-CGAN generates high-fidelity synthetic medical images that closely resemble real samples while improving the distinction between real and fake images. Second, a scalable TensorFlow Records (TFRecords)-based pipeline is implemented on Google Cloud Platform (GCP) to support efficient storage, loading, and processing of large-scale medical datasets. Third, a real-world COVID-19 medical image dataset comprising four disease classes is compiled and used to enhance diagnostic prediction. Experimental evaluation was conducted against several baseline generative models, including AC-GAN, WGAN, Pix2Pix, BigGAN, and CWGAN, with AC-GAN serving as the primary like-for-like baseline. Adaptive-CGAN improved classification accuracy from 93.75% ± 1.10% to 99.60% ± 0.40%, increased the Inception Score from 7.20 ± 0.28 to 8.47 ± 0.18 (+ 17.65%), reduced FID from 28.41 ± 1.35 to 26.13 ± 0.82 (− 8.02%), and reduced KID from 0.0156 ± 0.0021 to 0.0143 ± 0.0015 (− 8.33%). Adaptive-CGAN also reduced training time by 15.8% and CO₂ emissions by 9.5% compared with AC-GAN. Moreover, the GCP-based Adaptive-CGAN deployment emitted only 0.0023 kg CO₂, compared with 0.0047 kg CO₂ for the local setup, representing an approximately 50% reduction due to optimized cloud execution and the TFRecord-based pipeline. These results demonstrate that Adaptive-CGAN can improve diagnostic performance, synthetic image quality, computational efficiency, and environmental sustainability in AI-assisted healthcare.
W. Saber, A. El-Baz, R. Rizk· Cluster Computing· 0 citations