The "man vs. machine" battle in the medical laboratory field was fiercely staged on July 7, 2023, at the 2023 International Hematology Clinical and Translational Conference and the 12th Two Rivers Forum on Hematology and Oncology. German artificial intelligence competed against over a hundred experts in morphological testing from across the country, engaging in a live battle of blood cell morphology recognition. Ultimately, the German AI outperformed humans with a recognition speed of 2 seconds per image and a 100% accuracy rate.

The International Clinical and Translational Symposium on Hematology and the 12th Liangjiang Forum on Hematology and Oncology is hosted by the Chinese Medical Association Press and organized by the Hematology Medical Center of Xinqiao Hospital, Army Medical University. This conference invited top experts in the field of hematology research from home and abroad to share their latest research and the transformation of achievements, and also set up multiple specialized sub-forums. At the sub-forum "Clinical Application Forum of Artificial Intelligence Recognition of Blood Cells", Chongqing Defang Information Technology Co., LTD. 's national exclusive product "Peripheral Blood Image Analysis System DF-XX100S Hooke" was showcased on the spot and competed with over a hundred morphological examination experts from all over the country, recognizing 25 blood cell morphological images on the spot. After a tense and fierce competition, among over a hundred human contestants, two achieved full marks with all their images recognized correctly. The fastest one took 409 seconds, while the German AI peripheral blood image analysis system DF-XX100S Hooke emerged victorious with a recognition speed of 2 seconds per image and a 100% accuracy rate.

 

Highlights of the meeting

Blood cell morphology examination is the "gold standard" for diagnosing blood system diseases. The pathological basis of blood system diseases can be discovered through the observation of blood cell morphology under a microscope. However, the traditional manual microscopic examination method is a time-consuming and laborious physical task. Experienced laboratory personnel need at least 60 minutes to examine a sample under a microscope. Meanwhile, morphological examination is also an experience-based discipline, and the talent cultivation period is long. As a result, this important "gold standard" testing item is difficult to carry out in most grassroots hospitals. With the wide application of artificial intelligence technology in the medical field, the technology has become increasingly mature. By leveraging the deep learning capabilities of AI and its advantage of being free from human subjective limitations, it can effectively solve the predicament of manual microscopic examination, liberate human labor, and address issues such as difficult recognition, inaccurate recognition, and inconsistent standards. The DF-XX100S peripheral blood image analysis system by Hooke, based on artificial intelligence technology, has enhanced the diagnostic level and work efficiency of clinical tests. It can automatically classify and count 49 types of abnormal blood cells (red blood cells, white blood cells, platelets), and provide auxiliary diagnostic basis for blood diseases. It is a "good helper" for clinical blood cell morphology diagnosis and peripheral blood re-examination.

The outstanding performance of the peripheral blood image analysis system DF-XX100S Hooke in this man-machine battle has been recognized by many industry experts, proving its efficiency and accuracy in peripheral blood morphology examination. At present, this product has also been applied in many top-tier hospitals. In the future, it will be vigorously promoted and applied in hospitals at all levels across the country, making the screening of blood system diseases more efficient and accurate, and benefiting more patients.