Auxiliary Pneumonia Classification Algorithm Based on Pruning Compression.

Pneumonia infection is the leading cause of death in young children. The commonly used pneumonia detection method is that doctors diagnose through chest X-ray, and external factors easily interfere with the results. Assisting doctors in diagnosing pneumonia in patients based on deep learning methods can effectively eliminate similar problems. However, the complex network structure and redundant parameters of deep neural networks and the limited storage and computing resources of clinical medical hardware devices make it difficult for this method to use widely in clinical practice. Therefore, this paper studies a lightweight pneumonia classification network, CPGResNet50 (ResNet50 with custom channel pruning and ghost methods), based on ResNet50 pruning and compression to better meet the application requirements of clinical pneumonia auxiliary diagnosis with high precision and low memory. First, based on the hierarchical channel pruning method, the channel after the convolutional layer in the bottleneck part of the backbone network layer is used as the pruning object, and the pruning operation is performed after its normalization to obtain a network model with a high compression ratio. Second, the pruned convolutional layers are decomposed into original convolutions and cheap convolutions using the optimized convolution method. The feature maps generated by the two convolution parts are combined as the input to the next convolutional layer. Further, we conducted many experiments using pneumonia X-ray medical image data. The results show that the proposed method reduces the number of parameters of the ResNet50 network model from 23.7 M to 3.455 M when the pruning rate is 90%, a reduction is more than 85%, FIOPs dropped from 4.12G to 523.09 M, and the speed increased by more than 85%. The model training accuracy error remained within 1%. Therefore, the proposed method has a good performance in the auxiliary diagnosis of pneumonia and obtained good experimental results.

Characteristics of the Distance and Space Required for Knife Slashing Using a Motion Capture System.

OBJECTIVES: To investigate the relationship between the perpetrator's sex, victim's position and slashing location as well as anthropometric parameters on distance and space required for slashing, to provide the theoretical basis for the judgment of whether the crime scene was consistent with the criminal activity space. METHODS: The kinematics data of 12 male and 12 female subjects slashing the neck of standing and supine mannequins as well as the chest of the standing mannequins with a kitchen knife were obtained by using a 3D motion capture system. The relationship between the perpetrator's sex-victim's position, the perpetrator's sex-slashing location, and anthropometric parameters and the distance and space required for the slashing were analyzed by two-factor repeated measures ANOVA and Pearson correlation analysis respectively. RESULTS: Compared with slashing the neck of supine mannequins, the distance (L) and normalized L (l) of slashing the neck of standing mannequins were greater, while vertical distance (LVR) and normalized LVR (lVR) of the knife side were smaller. Compared with slashing the neck of standing mannequins, the L and l slashing the chest of standing mannequins were greater, while LVR and lVR were smaller. Horizontal distance (LHR) and normalized LHR (lHR) of the knife side in males were greater than that in females. Height and arm length were positively correlated with L, LHR, and LVR when striking the standing mannequins. CONCLUSIONS: When slashing the neck of supine or standing victims, the slashing distance is shorter and the slashing height is greater. Furthermore, the distance and space required for slashing are correlate with anthropometric parameters.