Accurate and quantitative pain assessment remains a major challenge in clinical medicine, especially for patients unable to verbalize discomfort. Conventional methods based on self-reports or clinician observation are subjective and inconsistent. This study introduces a novel automated facial pain assessment framework built on a dual-attention convolutional neural network (CNN) that achieves clinically calibrated, high-reliability performance and interpretability. The architecture combines multi-head spatial attention to localize pain-relevant facial regions with an enhanced channel attention block employing triple-pooling (average, max, and standard deviation) to capture discriminative intensity features. Regularization through label smoothing (α = 0.1) and AdamW optimization ensures calibrated, stable convergence. Evaluated on a clinically annotated dataset using subject-wise stratified sampling, the proposed model achieved a test accuracy of 90.19% ± 0.94%, with an average 5-fold cross-validation accuracy of 83.60% ± 1.55%. The model further attained an F1-score of 0.90 and Cohen’s κ = 0.876, with macro- and micro-AUCs of 0.991 and 0.992, respectively. The evaluation covers five pain classes (No Pain, Mid Pain, Moderate Pain, Severe Pain, and Very Pain) using subject-wise splits comprising 5840 total images and 1160 test samples. Comparative benchmarking and ablation experiments confirm each module’s contribution, while Grad-CAM visualizations highlight physiologically relevant facial regions. The results demonstrate a robust, explainable, and reproducible framework suitable for integration into real-world automated pain-monitoring systems. Inspired by biological pain perception mechanisms and human facial muscle responses, the proposed framework aligns with biomimetic sensing principles by emulating how localized facial cues contribute to pain interpretation.

Currently, advances in healthcare technologies are transforming medical diagnostics, particularly for data-driven disease detection. Acute lymphoblastic leukemia is a common and life-threatening blood cancer, especially prevalent in children. Although existing artificial intelligence-based diagnostic models have demonstrated promising accuracy, their black-box nature limits transparency, explainability, and clinical trust. Moreover, most explainable artificial intelligence approaches for leukemia diagnosis rely primarily on qualitative visual explanations and lack a unified quantitative mechanism to measure trust. This study presents a novel, trust-centered explainable deep-learning framework for automated acute lymphoblastic leukemia detection using 153 publicly available microscopic blood smear images and multiple transfer-learning models. Among the evaluated architectures, the fine-tuned EfficientNetB4 achieved high diagnostic performance, attaining an accuracy of 98.31%, and was selected as the base model of the proposed framework. Beyond diagnostic accuracy, this work introduces a novel unified interpretability score that quantitatively assesses model trustworthiness by integrating diagnostic performance, explanation robustness (evaluated using deletion and stability metrics), and clinician trust feedback into a single reliability measure. This quantitative trust formulation represents a distinct advancement over existing studies based on explainable artificial intelligence, which typically rely on isolated or qualitative explainability assessments. The framework further enhances transparency through visual and textual explanations generated using standard post hoc explainable methods and advanced fusion heatmaps, while human-centric evaluation is conducted using two clinician-based trust scales. Overall, this study provides a unified and transparent framework that jointly evaluates accuracy, explainability, and clinician trust, representing a step towards the clinical adoption of trustworthy artificial intelligence driven leukemia diagnostic systems.

Deep learning has transformed medical image analysis, achieving high accuracy in classification, segmentation, and diagnosis. However, the substantial computational and energy demands of advanced models, such as large convolutional neural networks (CNNs) and multimodal vision-language systems, hinder their deployment on resource-constrained edge devices like smartphones, especially in mobile health. This challenge is critical in low-resource, bandwidth-limited settings with limited high-performance computing access, exacerbating healthcare disparities. This paper introduces MobileNet-Lite-Health (MLH), a lightweight CNN framework optimized via quantization, pruning, and knowledge distillation for sustainable medical imaging inference.We assess MLH on MedMNIST (PathMNIST subset) and HAM10000 datasets, representing benchmark and clinical contexts, respectively. Key contributions include: (1) a carbon-aware evaluation pipeline integrating CodeCarbon for CO2 tracking and real-time battery profiling on devices like the Google Pixel 6; (2) comparisons with ResNet-18 and MobileNetV2, showing 74% energy reduction and 76% CO2 decrease per inference with a 2% accuracy drop; and (3) scaling simulations projecting approximately 1.1 g annual CO2 savings per clinic for 50,000 screenings. MLH serves as a green alternative to energy-intensive models like GPT-4V and MedPaLM, unsuitable for edge use. Experimental results confirm MLH’s effectiveness, with 84.7% accuracy on PathMNIST and 80.1% on HAM10000, plus low latency (65 ms) and memory use (95 MB). This work pioneers carbon-aware edge AI, promoting sustainable, accessible healthcare in resource-scarce environments by balancing efficiency, utility, and environmental impact.

Accurate and quantitative pain assessment remains a major challenge in clinical medicine, especially for patients unable to verbalize discomfort. Conventional methods based on self-reports or clinician observation are subjective and inconsistent. This study intro-duces a novel automated facial pain assessment framework built on a dual-attention Convolutional Neural Network (CNN) that achieves Medical-Grade reliability and inter-pretability. The architecture combines multi-head spatial attention to localize pain-relevant facial regions with an enhanced channel-attention block employing triple pooling (average, maximum, and standard-deviation) to capture discriminative intensity features. Regularization through label smoothing (α = 0.1) and AdamW optimization ensures calibrated, stable convergence. Evaluated on a clinically annotated dataset using subject wise stratified sampling, the proposed model achieved a test accuracy of 90.19 % ± 0.94 %, with an average 5-fold cross-validation accuracy of 83.60 % ± 1.55 %. The model further attained an F1-score of 0.90 and Cohen’s κ = 0.876, with macro and micro-AUCs of 0.991 and 0.992, respectively. The evaluation covers five pain classes (No Pain, Mid Pain, Moderate Pain, Severe Pain, and Very Pain) using subject-wise splits comprising 5,840 total images and 1160 test samples. Comparative benchmarking and ablation experiments confirm each module’s contribution, while Grad-CAM visualizations highlight physio-logically relevant facial regions. The results demonstrate a robust, explainable, and re-producible framework suitable for integration into real-world automated pain-monitoring systems.