Brain tumors require precise segmentation for diagnosis and treatment plans due to their complex morphology and heterogeneous characteristics. While MRI-based automatic brain tumor segmentation technology reduces the burden on medical staff and provides quantitative information, existing methodologies and recent models still struggle to accurately capture and classify the fine boundaries and diverse morphologies of tumors. In order to address these challenges and maximize the performance of brain tumor segmentation, this research introduces a novel SwinUNETR-based model by integrating a new decoder block, the Hierarchical Channel-wise Attention Decoder (HCAD), into a powerful SwinUNETR encoder. The HCAD decoder block utilizes hierarchical features and channel-specific attention mechanisms to further fuse information at different scales transmitted from the encoder and preserve spatial details throughout the reconstruction phase. Rigorous evaluations on the recent BraTS GLI datasets demonstrate that the proposed SwinHCAD model achieved superior and improved segmentation accuracy on both the Dice score and HD95 metrics across all tumor subregions (WT, TC, and ET) compared to baseline models. In particular, the rationale and contribution of the model design were clarified through ablation studies to verify the effectiveness of the proposed HCAD decoder block. The results of this study are expected to greatly contribute to enhancing the efficiency of clinical diagnosis and treatment planning by increasing the precision of automated brain tumor segmentation.

Small object detection in remote sensing imagery is a challenging task due to the small size of targets, complex background, and low contrast, which makes achieving high precision difficult. To enhance the accuracy of detection, this study proposes a novel oriented object detection model with three significant innovations: Firstly, a lightweight feature extraction network is designed to achieve efficient feature representation at a reduced computational cost, which is particularly effective for the recognition of small targets in remote sensing imagery. Secondly, a Feature-Focused Channel Attention (FFCA) is introduced that enhances the model’s ability to focus on small target areas by combining spatial and channel attention, enhancing the model’s capacity to capture and represent features more effectively. Lastly, an attention-guided multi-scale feature fusion module is proposed to integrate features from different levels, which substantially boosts the model’s ability to accurately detect small-scale objects, especially in remote sensing scenarios with vast fields of view and complex backgrounds. The experimental outcomes validate that our model achieves the best detection performance on two benchmark public datasets for remote sensing imagery, confirming its effectiveness and practicality in remote small object detection tasks.