Bio-net dataset: AI-based diagnostic solutions using peripheral blood smear images.

Manual evaluation of peripheral blood smears remains a standard in hematology. Automated complete blood count tests often precede microscopic confirmation. This process demands skilled technicians and takes several hours. Recent advances in deep learning have shown promise in reducing manual workloads. However, training AI models requires high-quality annotated datasets. Prior research has shown that automated systems can match human accuracy in some cases. No prior work had resolved the need for a large-scale dataset specific to healthy individuals. This gap motivated the creation of a new dataset to improve AI-based diagnostics. The Bio-Net dataset aims to fill this void in biomedical research.

The primary aim is to introduce a new annotated dataset for blood cell analysis. This dataset supports AI-based detection and counting of blood cells. The study also aims to demonstrate how AI can streamline diagnostic workflows. Blood cell classification is a core challenge in hematology. The Bio-Net dataset is designed to enhance automated diagnostic tools. A secondary goal is to provide a specialized subset for white blood cell classification. The dataset's structure allows for comparison with existing public datasets. This work addresses the need for better data in AI-driven diagnostics.

The Bio-Net dataset includes annotated images of red blood cells, white blood cells, and platelets. Images were obtained from peripheral blood smears of healthy individuals. A modified YOLO algorithm was trained on this dataset for object detection. The model was tested on white blood cell classification tasks specifically. The dataset includes a mini-version for specialized image processing tasks. Comparisons were made with other publicly available datasets to assess performance. The dataset's structure allows for both detection and classification tasks. This approach enables benchmarking of AI models in hematology.

The Bio-Net dataset contains a large number of annotated blood cell images. The dataset includes both full-scale and mini-versions for different tasks. The modified YOLO algorithm successfully detected and classified blood cells. Performance metrics showed strong agreement with existing public datasets. The model achieved high accuracy in white blood cell classification. The dataset's structure supports both detection and counting tasks. Results suggest potential for broader applications in biomedical research. These findings highlight the dataset's value in advancing AI diagnostics.

The Bio-Net dataset provides a valuable resource for AI-based blood cell analysis. The dataset's structure supports both detection and classification tasks. The modified YOLO algorithm demonstrated strong performance in cell classification. These results suggest the dataset's potential for broader diagnostic applications. The study highlights the importance of annotated datasets in AI development. The dataset's availability may improve diagnostic workflows in hematology. The findings align with the authors' stated goals of advancing biomedical research. This work contributes to the growing field of AI in medical diagnostics.