Transformative Deep Neural Network Approaches in Kidney Ultrasound Segmentation: Empirical Validation with an Annotated Dataset

Kidney ultrasound (US) images are primarily employed for diagnosing different renal diseases. Among them, one is renal localization and detection, which can be carried out by segmenting the kidney US images. However, kidney segmentation from US images is challenging due to low contrast, speckle noise, fluid, variations in kidney shape, and modality artifacts. Moreover, well-annotated US datasets for renal segmentation and detection are scarce. This study aims to build a novel, well-annotated dataset containing 44,880 US images. In addition, we propose a novel training scheme that utilizes the encoder and decoder parts of a state-of-the-art segmentation algorithm. In the pre-processing step, pixel intensity normalization improves contrast and facilitates model convergence. The modified encoder–decoder architecture improves pyramid-shaped hole pooling, cascaded multiple-hole convolutions, and batch normalization. The pre-processing step gradually reconstructs spatial information, including the capture of complete object boundaries, and the post-processing module with a concave curvature reduces the false positive rate of the results. We present benchmark findings to validate the quality of the proposed training scheme and dataset. We applied six evaluation metrics and several baseline segmentation approaches to our novel kidney US dataset. Among the evaluated models, DeepLabv3+ performed well and achieved the highest dice, Hausdorff distance 95, accuracy, specificity, average symmetric surface distance, and recall scores of 89.76%, 9.91, 98.14%, 98.83%, 3.03, and 90.68%, respectively. The proposed training strategy aids state-of-the-art segmentation models, resulting in better-segmented predictions. Furthermore, the large, well-annotated kidney US public dataset will serve as a valuable baseline source for future medical image analysis research. Graphic Abstract The graphic abstract for this research study visually encapsulates the key contributions and innovations: Dataset creation Developed WD-KUS dataset (44,880 US images). Aims to standardize US segmentation benchmarks and simplify US interpretation efforts. Automatic kidney segmentation framework Demonstrated a practical framework for segmenting whole kidneys from low-quality US images. Integrated various encoder–decoder models and a unique training strategy. Addressed challenges like shadows, internal fluid, and blurring. Training and post-processing strategies Introduced effective training strategies (pre-processing, networks, learning rate). Utilized post-processing techniques, including concave curvature assessment. Improved segmentation accuracy and generalizability. Auxiliary function for abnormal segmentation Proposed an auxiliary function to distinguish normal and abnormal kidney segmentation. Based on the concept that normal segmentation has few concave corners, while abnormal segmentation has many. Quantitative and qualitative enhancement Enhanced segmentation results quantitatively (six metrics). Showcased qualitative improvement over baseline methods. Validation with state-of-the-art networks Validated the approach using modified state-of-the-art segmentation neural networks. Demonstrated the effectiveness of the WD-KUS dataset in validation. The research significantly contributes to the field by providing a comprehensive dataset, an advanced segmentation framework, and innovative strategies for training and post-processing, resulting in improved kidney segmentation accuracy and applicability.