"IEEE BioCAS 2025 Grand Challenge on Pediatric Leukemia Diagnosis"

"The AI recognition of bone marrow\/blood cell images has become one of the important research directions in the field of medical imaging. By analyzing bone marrow images, doctors can quickly and accurately diagnose and monitor a range of blood diseases such as leukemia, anemia, and others. However, traditional bone marrow image analysis usually requires doctors to spend a lot of time and effort, and the results are subject to subjective factors, leading to issues such as inconsistent diagnoses.In recent years, with the development of artificial intelligence and machine learning technologies, AI recognition technology based on deep learning for bone marrow images has gradually matured. A large body of research indicates that using deep learning algorithms can effectively automatically identify abnormal cells, blood components, and structural features in bone marrow images, thereby assisting doctors in diagnosis and treatment decisions.However, despite significant progress in deep learning for bone marrow image recognition, there are still some challenges. These include difficulties in dataset annotation, insufficient generalization ability of models, and limitations in hardware resources. Especially when deploying large deep learning models on mobile medical devices and edge computing platforms, the limited availability of hardware resources becomes a constraining factor.In order to further promote the development and application of bone marrow image AI recognition technology, we will hold a challenge competition for bone marrow cell image AI recognition. This event aims to invite researchers, scholars, and medical experts from around the world to jointly explore and innovate bone marrow image AI recognition technology, improve the accuracy and efficiency of bone marrow disease diagnosis. Through this challenge competition, we hope to leverage artificial intelligence technology to bring faster, more accurate, and reliable diagnostic solutions to the field of medical imaging, provide better auxiliary tools for clinical doctors, and promote the development of healthcare."

骨髓/血细胞图像的AI识别现已成为医学影像领域的重要研究方向之一。通过分析骨髓图像，临床医师能够快速且精准地诊断并监测白血病、贫血等多种血液疾病。然而传统骨髓图像分析往往需要医师耗费大量时间与精力，且分析结果易受主观因素影响，易出现诊断结果不一致等问题。近年来，随着人工智能（Artificial Intelligence）与机器学习技术的发展，基于深度学习的骨髓图像AI识别技术逐渐趋于成熟。大量研究表明，运用深度学习算法可有效实现骨髓图像中异常细胞、血液成分及结构特征的自动识别，进而辅助医师开展诊断与治疗决策。然而尽管骨髓图像识别的深度学习技术已取得长足进展，仍面临诸多挑战：包括数据集标注难度大、模型泛化能力不足以及硬件资源受限等问题。尤其是在将大型深度学习模型部署至移动医疗设备与边缘计算平台时，硬件资源有限便成为制约其应用的关键因素。为进一步推动骨髓图像AI识别技术的发展与应用，我们将举办骨髓细胞图像AI识别挑战赛。本次赛事旨在邀请全球范围内的科研人员、学者与医学专家，共同探索并创新骨髓图像AI识别技术，提升骨髓疾病诊断的精准度与效率。通过本次挑战赛，我们期望借助人工智能技术，为医学影像领域带来更快速、精准且可靠的诊断解决方案，为临床医师提供更优质的辅助工具，进而推动医疗健康事业的发展。