Efficient and explainable vehicle damage inspection is essential due to the increasing complexity and volume of vehicular incidents. Traditional manual inspection approaches are not time-effective, prone to human error, and lead to inefficiencies in insurance claims and repair workflows. Existing deep learning methods, such as CNNs, often struggle with generalization, require large annotated datasets, and lack interpretability. This study presents a robust and interpretable deep learning framework for vehicle damage classification, integrating Vision Transformers (ViTs) and ensemble detection strategies. The proposed architecture employs a RetinaNet backbone with a ViT-enhanced detection head, implemented in PyTorch using the Detectron2 object detection technique. It is pretrained on COCO weights and fine-tuned through focal loss and aggressive augmentation techniques to improve generalization under real-world damage variability. The proposed system applies the Weighted Box Fusion (WBF) ensemble strategy to refine detection outputs from multiple models, offering improved spatial precision. To ensure interpretability and transparency, we adopt numerous explainability techniques—Grad-CAM, Grad-CAM++, and SHAP—offering semantic and visual insights into model decisions. A custom vehicle damage dataset with 4500 images has been built, consisting of approximately 60% curated images collected through targeted web scraping and crawling covering various damage types (such as bumper dents, panel scratches, and frontal impacts), along with 40% COCO dataset images to support model generalization. Comparative evaluations show that Hybrid ViT-RetinaNet achieves superior performance with an F1-score of 84.6%, mAP of 87.2%, and 22 FPS inference speed. In an ablation analysis, WBF, augmentation, transfer learning, and focal loss significantly improve performance, with focal loss increasing F1 by 6.3% for underrepresented classes and COCO pretraining boosting mAP by 8.7%. Additional architectural comparisons demonstrate that our full hybrid configuration not only maintains competitive accuracy but also achieves up to 150 FPS, making it well suited for real-time use cases. Robustness tests under challenging conditions, including real-world visual disturbances (smoke, fire, motion blur, varying lighting, and occlusions) and artificial noise (Gaussian; salt-and-pepper), confirm the model’s generalization ability. This work contributes a scalable, explainable, and high-performance solution for real-world vehicle damage diagnostics.

Type 2 diabetes mellitus represents a global public health challenge, with rising prevalence driven by complex interactions between lifestyle factors, health literacy, cultural beliefs, and demographic characteristics. Despite extensive research, few studies have systematically examined the interplay between behavioural determinants, cognitive awareness, misconceptions, and computational prediction models within integrated frameworks. This systematic review synthesises evidence on multidimensional risk factors for type 2 diabetes, examining lifestyle behaviours, health literacy and awareness, cultural misconceptions, family history, and the application of machine learning approaches in risk prediction and behavioural profiling. A systematic search was conducted across major databases including PubMed, Scopus, and Web of Science. Studies published between 2008 and 2025 examining lifestyle factors, health literacy, awareness, misconceptions, family history, and computational approaches related to type 2 diabetes risk were included. Quality assessment was performed, and data were synthesised narratively across five thematic domains. Thirteen studies met inclusion criteria, revealing a fragmented evidence base. While individual domains showed strong associations (e.g., sedentary behaviour with diabetes risk, health literacy with preventive behaviour, and cultural misconceptions with treatment adherence), no study successfully integrated behavioural, cognitive, and computational factors within a single predictive framework. Health literacy and awareness significantly influenced preventive behaviours, while cultural misconceptions impeded effective disease management. Family history emerged as a consistent non-modifiable risk factor. Machine learning models demonstrated high predictive accuracy but often lacked behavioural and cognitive variables, limiting their clinical applicability. Few studies integrated multiple dimensions simultaneously. This review highlights critical gaps in holistic diabetes risk assessment. Future research should develop integrated frameworks combining behavioural profiling, cognitive assessments, and explainable artificial intelligence to enable personalised prevention strategies and improve clinical decision making.