BACKGROUND:Knee osteoarthritis is a common degenerative disease that significantly impacts patients'quality of life and increases the societal healthcare burden.Early and accurate diagnosis of knee osteoarthritis is crucial for the treatment and prognosis of patients.Traditional diagnostic methods are not only subjective and time-consuming but also do not guarantee consistently high accuracy.OBJECTIVE:To develop an automatic diagnostic method for knee osteoarthritis based on deep learning,utilizing deep learning networks to improve diagnostic accuracy and efficiency.METHODS:A new network model,YOLOV8-ViT,was proposed by replacing the backbone network of YOLOv8n with the Efficient-ViT network and incorporating attention mechanisms for the automatic identification and classification of X-ray images of knee osteoarthritis.The experimental dataset included 5 078 X-ray images of patients with knee osteoarthritis obtained from the Third Affiliated Hospital of Guangzhou University of Chinese Medicine.Three imaging physicians annotated the sites of knee osteoarthritis and classified them according to the Kellgren-Lawrence grading standard using Labelme software,and the results were combined.The evaluation indicators used in this study included Precision,F1 score,mean average precision(mAP),Recall,val/box_loss,val/cls_loss,and val/dfl_loss.RESULTS AND CONCLUSION:The experimental results showed that the YOLOV8-ViT model outperformed the YOLOv5n,YOLOv8n,and YOLOv9n models in terms of precision,mAP50,mAP50-95,F1 score,and Recall,while lowering val/box_loss,val/cls_loss,and val/dfl_loss by 0.496,0.45,and 0.523;1.037,0.305,and 0.728;and 0.267,0.654,and 0.854,respectively.These experimental data validate that this model has high detection accuracy.

BACKGROUND:Knee osteoarthritis is a common degenerative disease that significantly impacts patients'quality of life and increases the societal healthcare burden.Early and accurate diagnosis of knee osteoarthritis is crucial for the treatment and prognosis of patients.Traditional diagnostic methods are not only subjective and time-consuming but also do not guarantee consistently high accuracy.OBJECTIVE:To develop an automatic diagnostic method for knee osteoarthritis based on deep learning,utilizing deep learning networks to improve diagnostic accuracy and efficiency.METHODS:A new network model,YOLOV8-ViT,was proposed by replacing the backbone network of YOLOv8n with the Efficient-ViT network and incorporating attention mechanisms for the automatic identification and classification of X-ray images of knee osteoarthritis.The experimental dataset included 5 078 X-ray images of patients with knee osteoarthritis obtained from the Third Affiliated Hospital of Guangzhou University of Chinese Medicine.Three imaging physicians annotated the sites of knee osteoarthritis and classified them according to the Kellgren-Lawrence grading standard using Labelme software,and the results were combined.The evaluation indicators used in this study included Precision,F1 score,mean average precision(mAP),Recall,val/box_loss,val/cls_loss,and val/dfl_loss.RESULTS AND CONCLUSION:The experimental results showed that the YOLOV8-ViT model outperformed the YOLOv5n,YOLOv8n,and YOLOv9n models in terms of precision,mAP50,mAP50-95,F1 score,and Recall,while lowering val/box_loss,val/cls_loss,and val/dfl_loss by 0.496,0.45,and 0.523;1.037,0.305,and 0.728;and 0.267,0.654,and 0.854,respectively.These experimental data validate that this model has high detection accuracy.

Wall filtering is a key technology in ultrasound color flow imaging system. Without efficient suppression of wall signal originating from stationary and moving tissue, low velocity blood flow cannot be measured, and the estimates of higher velocities will have a large bias. Among the various wall filters, the projection-initialized infinite impulse response (IIR) wall filter has narrow transition bandwidth and high stopband suppression ratio; it is superior to other wall filters. At present, the related literatures are only limited to theoretical research on this kind of filter, and no feasible design and implementation methods are reported. In this paper, a new design and implementation scheme for the projection-initialized filter is proposed, which transforms the filtering process to matrix multiplications. The proposed scheme is realized on programmable logic devices. Experimental results show that it is a simple and effective implementation method for projection-initialized IIR filter, and it is superior to conventional wall filters.
Algorithms ; Blood Flow Velocity ; physiology ; Blood Vessels ; diagnostic imaging ; Computer Simulation ; Equipment Design ; Image Interpretation, Computer-Assisted ; Signal Processing, Computer-Assisted ; Ultrasonography, Doppler, Color ; instrumentation ; methods

OBJECTIVE: To explore a set of quantitative methods to determine the position of the facial nerve based on three-dimensional CT reconstruction of temporal bone structures on personal computer, which can provide a series of important parameters for ear and the lateral skull base surgery. METHOD: The internal structures of temporal bone were reconstructed based on a set of axial CT images of adult patients, the complicated relationship and their morphologic characteristics were clearly presented by using Able Software 3D-DOCTOR. The precise measurement of some parameters between facial nerve and its adjacent structures could easily be processed with the software. Based on all obtained data, the relationship of facial nerve and its adjacent structures were effectively summarized and analyzed. RESULT: Three-dimensional images, including the facial nerve, tympanic anulus, auditory ossicles, chochleariform process, pyramidal eminence, internal auditory, the cochlea, semicircular canal, jugular fossa and carotid artery in the temporal bone, were reconstructed. Some parameters obtained from measuring the distance or angle between the facial nerve and its adjacent structures in the three-dimensional models had some extent regularity, which were benefit to design surgical approach and determine the position of facial nerve during relevant operation. CONCLUSION CT 3D reconstruction can accurately display the detailed internal structures anatomy of the temporal bone and their quantitative spatial relationships.
Adult ; Facial Nerve ; anatomy & histology ; diagnostic imaging ; Female ; Humans ; Imaging, Three-Dimensional ; Male ; Middle Aged ; Temporal Bone ; diagnostic imaging ; Tomography, X-Ray Computed ; methods

Objective To establish optimum processing method for Pollen Typhae Carbonisatus. Method Processing method was studied by orthogonal test and the total flavones were determined by HPLC. Conditions of HPLC used to determine the total flavones were: Waters Nove-Park C18 (150 mm?3.9 mm, 4 ?m), mobile phase: MeOH-THF-0.05%TFT (16∶24∶60), flow rate: 0.8 mL/min, column temperature: 30 ℃, detection wavelength: 360 nm. Results The optimum processing method was skir-baked for 8 min at 210 ℃. Conclusion The optimized processing method is available for the processing of Pollen Typhae Carbonisatus.