Skin cancer is one of the widespread diseases among existing cancer types. More importantly, the detection of lesions in early diagnosis has tremendously attracted researchers’ attention. Thus, artificial intelligence (AI)-based techniques have supported the early diagnosis of skin cancer by investigating deep-learning-based convolutional neural networks (CNN). However, the current methods remain challenging in detecting melanoma in dermoscopic images. Therefore, in this paper, we propose an ensemble model that uses the vision of both EfficientNetV2S and Swin-Transformer models to detect the early focal zone of skin cancer. Hence, we considerthat the former architecture leads to greater accuracy, while the latter model has the advantage of recognizing dark parts in an image. We have modified the fifth block of the EfficientNetV2S model and have included the Swin-Transformer model. Our experiments demonstrate that the constructed ensemble model has attained a higher level of accuracy over the individual models and has also decreased the losses as compared to traditional strategies. The proposed model achieved an accuracy score of 99.10%, a sensitivity of 99.27%, and a specificity score of 99.80%.

Leukemia is a deadly disease caused by the overproduction of immature white blood cells (WBS) in the bone marrow. If leukemia is detected at the initial stages, the chances of recovery are better. Typically, morphological analysis for the identification of acute lymphoblastic leukemia (ALL) is performed manually on blood cells by skilled medical personnel, which has several disadvantages, including a lack of medical personnel, sluggish analysis, and prediction that is dependent on the medical personnel’s expertise. Therefore, we proposed the Multi-Attention EfficientNetV2S and EfficientNetB3 state-of-the-art deep learning architectures using transfer learning-based fine-tuning approach to distinguish the normal and blast cells from microscopic blood smear images that both are pretrained on large-scale ImageNet database. We simply modified the last block of both models and added additional layers to both models. After including this Multi-Attention Mechanism, it not only reduces the model’s complexities but also generalizes its network quite well. By using the proposed technique, the accuracy has improved and the overall loss is also minimized. Our Multi-Attention EfficientNetV2S and EfficientNetB3 models achieved 99.73% and 99.25% accuracy, respectively. We have further compared the proposed model’s performance to other individual and ensemble models. Upon comparison, the proposed model outclassed the existing literature and other benchmark models, thus proving its efficiency.

Due to the limited availability of training data, the diverse shapes of brain tumors among different patients, inter-class similarity, and intra-class variation, achieving high recognition accuracy and speed in deep learning-based brain tumor recognition remains challenging. To address these issues, we propose a Dual-Branch Ensemble and Gated Global-Local Attention network based on EfficientNetV2S and ConvNeXt (GGLA-NeXtE2NET) to improve identification accuracy and model interpretability. For inter-class and intra-class problems, we designed a Gated Global-Local Attention (GGLA) mechanism that captures dependency information of query points in both horizontal and vertical directions, thereby obtaining global information indirectly. Simultaneously, local information is captured through multiple convolutions with a gating layer. The gating mechanism within the GGLA dynamically balances the contributions of global and local information, enabling the model to adaptively focus on the most relevant features for accurate classification. Furthermore, we introduce a dual-branch ensemble network to address the issue of image variety. This network uses two branches to extract image features at different resolutions for fusion, thereby expanding the network receptive field. Additionally, we utilized an Enhanced Super-Resolution Generative Adversarial Network (ESRGAN) to generate images that balance MRI data and implemented multiple preprocessing techniques to tackle inherent noise in MRI images. These techniques enhance the clarity of MRI images while preserving essential details. This results in a clear improvement in the identification of tumor boundaries, crucial for accurate surgical planning and treatment strategies. We evaluated GGLA-NeXtE2NET on 3-class and 4-class brain tumor datasets and achieved 99.06%, and 99.62% overall accuracy on both datasets respectively.

The accurate diagnosis of skin cancer remains challenging due to factors like intra-class variation and inter-class similarity in dermoscopic images. Despite their sophistication, existing methods struggle to generalize well across diverse cases, and their computational complexity often limits real-world applicability. To address these challenges, we introduce a novel dual-path architecture in the Efficient Global-Vision Attention-Net (EG-VAN) for early detection of skin cancer. The dual-path design uniquely combines (1) EfficientNetV2S, which processes multi-scale features with high efficiency, and (2) a modified ResNet50 enhanced with a Spatial-Context Group Attention (SCGA) module and Non-Local Block. The SCGA module captures local-contextual relationships in horizontal and vertical directions to focus on subtle lesion details, directly addressing intra-class variation, while the Non-Local Block captures long-range dependencies to resolve inter-class similarity. This synergistic dual-path fusion ensures comprehensive feature extraction, enabling EG-VAN to outperform single-path architectures. Further, we introduce a Multi-Scale Feature Fusion (MFF) module to aggregate hidden features across scales, improving generalization across skin cancer types. Additionally, our pioneering color balancing method, inspired by Gray World and Retinex theory, recalibrates color distribution to highlight lesion regions, enhancing the dual-path network’s input quality. We validate our EG-VAN through extensive experiments on the 9-class skin cancer dataset. Our experiments show improvement in performance by achieving an accuracy of 98.20%, a recall of 95.74%, and an F1-Score of 96.68%.