Skin cancer is a potentially fatal disease that requires early and accurate diagnosis to improve patient outcomes. Deep learning has shown promise in automating skin lesion classification; however, many existing models suffer from limited interpretability, class imbalance, and irrelevant feature extraction. This study introduces CBAM-Xception, an explainable attention-guided deep learning model designed to enhance classification performance by focusing on clinically relevant lesion features. The model integrates a pretrained Xception backbone with a Convolutional Block Attention Module (CBAM) to highlight discriminative regions while suppressing background noise. CLAHE was applied to enhance contrast in dermoscopic images, and geometric and color augmentation were used to address class imbalance in the HAM10000 (seven classes) and ISIC 2019 (nine classes) datasets. The evaluation was performed using a standard training, validation, and test split. The first 50 layers of Xception were frozen before fine-tuning. Grad-CAM++ visualizations confirmed the focus of the model on key lesion areas. The proposed model achieved superior performance compared to the MobileNet and EfficientNet baselines, attaining accuracies of 98.62% (AUC: 0.9997) on HAM10000 and 93.66% (AUC: 0.9939) on ISIC 2019, demonstrating its strong effectiveness. However, the model's performance may depend on dataset-specific characteristics and computational resources, which could limit its generalizability to unseen clinical environments. By combining high accuracy, interpretability, and robustness to class imbalances, CBAM-Xception provides a reliable solution for automated skin cancer diagnosis.
The Saudi Public Investment Fund’s coordinated acquisition of four Saudi Pro League clubs, combined with transfer expenditure exceeding £800 million in a single window, has exposed a structural collapse at the heart of global football governance: existing financial sustainability frameworks were designed to discipline clubs, not sovereign states. This paper advances a doctrinal argument in three movements. First, it maps the jurisdictional void within the global regulatory hierarchy from FIFA through the AFC to the SPL, demonstrating that UEFA’s Club Licensing and Financial Sustainability Regulations, widely regarded as the gold standard of financial control, are structurally inapplicable to sovereign actors deliberately operating outside their jurisdictional reach. Second, it anatomizes the legal mechanisms through which the PIF exploits this void: corporate separation, sovereign immunity doctrine, and systematic jurisdictional arbitrage between continental confederations operating under irreconcilably different regulatory baselines. Third, it evaluates the doctrinal limits of available reform, engaging critically with the Meca-Medina proportionality framework, the structural weaknesses of CAS enforcement architecture, and the fundamental inadequacy of lex sportiva as a regulatory response to loss-insensitive state capital. Against this diagnosis, the paper proposes a tripartite governance blueprint: mandatory sovereignty waivers embedded within FIFA’s transfer regulatory framework, strategic deployment of the EU Foreign Subsidies Regulation as an extraterritorial enforcement backstop, and a Reverse-Assurance Guarantee mechanism requiring sovereign beneficial owners to submit to independent financial oversight as a condition of participation in the global transfer market.
This article interrogates the Westphalian requirement of territoriality as a prerequisite for statehood, proposing a reconstituted model of non-territorial legal personality to address the existential threat of climate-induced displacement. Utilizing a TWAIL-informed critique, the study identifies the geographic trap within the Montevideo Convention as a colonial artifact that facilitates the legal erasure of vanishing island nations and displaced polities. By decoupling legal subjectivity from physical soil, the research reinterprets Article 1 of the ICCPR to frame the continuity of the polity as a functional necessity for the protection of collective human rights. Drawing on precedents from the ICJ and UNGA regarding governments-in-exile and observer status, the paper argues for the recognition of Necessary Subjects within international law. It provides a doctrinal blueprint for a post-territorial legal order, ensuring that the erosion of sovereign land does not result in the termination of the human rights regime for the dispossessed.
Accurate sheep breed classification is critical for modern livestock management, supporting decisions on breeding, productivity, and profitability in the farming industry. Traditional methods based on visual inspection or body measurements are often subjective, time-consuming, and unsuitable for large-scale operations. To address these challenges, this study proposes SheepFormers, a Vision Transformer (ViT)-based framework for automatic sheep breed identification, specifically designed to operate effectively even with small datasets. A balanced dataset of 1,680 sheep face images representing four breeds was preprocessed and used to evaluate multiple transformer variants, including ViT-Google, ViT-MAE, ViT-VAN, BEiT, ViT-ResNet50, and DiT. Through systematic hyperparameter optimization and ablation studies covering epochs, batch size, learning rate, positional encoding, patch size, data augmentation, and 5-fold cross-validation, the ViT-Google model achieved the best performance with 98.21
Skin diseases pose a significant public health challenge, impacting millions of people globally with a wide range of phenomena, from mild conditions to life-threatening. In skin diseases, people are well known with an autoimmune disorder, which results from abnormal responses of the adaptive body’s immune system. Vitiligo is also a part of chronic autoimmune disorder, where a particular area of human skin loses its original colors, and people are affected in different ratios worldwide. This study investigates the effectiveness of deep learning models for detecting and segmenting vitiligo lesions in dermatological images. Two models, VGG19-ResSE and YOLOv8, were evaluated for their performance. The VGG19-ResSE model achieved an accuracy of 95.5
