The Study Of Identification Of Modes And On Mechanisms Of Head Injury Using The Combined Digital Technique

Background:Craniocerebral injury is the most common mechanical injury in the practice of forensic medicine,ranking first among violent deaths.When the human head is subj ected to blunt external forces such as acceleration,deceleration,squeezing,traction,etc.,it can cause damage to the scalp,skull and brain.Due to the complexity of the craniocerebral structure and the ambiguity of the current craniocerebral injury mechanism,it forms a technical bottleneck for judging the injury mode of craniocerebral injury in practice.Traditional forensic medicine mostly relies on the morphological characteristics of the injury site and expert experience in the identification of craniocerebral injury.The identification technology is single.The expert experience based on induction is difficult to obtain scientific argumentation,and there is a certain subjectivity.Therefore,it is necessary to seek new methods for the identification of injury modes and injury mechanisms of brain injury,to clarify the dose-effect relationship between external force,injury mechanism and injury results,reduce the interference of subjective opinions,increase the reproducibility of evidence,Objectivity and scientific.In recent years,virtual anatomy technology has been gradually used as an auxiliary means of anatomy in the identification of craniocerebral injuries.Its three-dimensional and intuitive display has significant advantages.The application of finite element method and other digital simulation methods to carry out craniocerebral injury biomechanical research has also achieved initial results With the advancement of computer technology,the great potential of artificial intelligence learning technology in image recognition also provides new technical means for the identification of traumatic methods of brain injury.Methods:(1)Retrospectively screen cases of craniocerebral injury based on case CT data,including typical blow injuries and falls.All cases should include detailed injury process,injury characteristics,degree of injury,imaging data,injury material,injury method,etc.,and comprehensively establish a brain injury case database、Use statistical methods to analyze the differences in the location of the force and the type(and distribution)of craniocerebral injuries under different injury modes,and explore the characteristics and general rules of craniocerebral injury under different injury modes.(2)The LVRNet residual network model for automatic identification of craniocerebral injuries based on deep learning technology can complete the classification of craniocerebral injury patterns to a certain extent.This model is more efficient in identifying accelerated craniocerebral injuries than decelerating cranial injuries.Brain damage:The algorithm is similar to the human eye in the pseudo-color image generated by the damage information.(3)The parametric study of accelerated and decelerated craniocerebral injuries based on finite element simulation technology found that when simulating acceleration and deceleration craniocerebral injuries,in addition to different stress propagation directions and different response trends,the acceleration and deceleration brain inj uries The biggest difference is the change of intracranial pressure at the impact point and the counterpoint:the intracranial pressure at the impact site of deceleration injury has a large value and a large range of changes,and the intracranial pressure at the counter site has a greater range of changes and a longer duration.Results:(1)Comprehensive analysis of the CT data of the case database found that:accelerated craniocerebral injury most often occurs in the temporal region,and the types of injury include scalp injury,skull fracture,impact brain contusion,and acute epidural hematoma.The main feature of craniocerebral deceleration injury is that it often occurs in the occiput,and the types of injury include scalp injury,subarachnoid hemorrhage,hedging brain contusion,and subdural hematoma.(2)The LVRNet residual network model for automatic identification of craniocerebral injuries based on deep learning technology can complete the classification of craniocerebral injury patterns to a certain extent.This model is more efficient in identifying accelerated craniocerebral injuries than decelerating cranial injuries.Brain damage:The algorithm is similar to the human eye in the pseudo-color image generated by the damage information.(3)The parametric study of accelerated and decelerated craniocerebral injuries based on finite element simulation technology found that when simulating acceleration and deceleration craniocerebral injuries,in addition to different stress propagation directions and different response trends,the acceleration and deceleration brain injuries The biggest difference is the change of intracranial pressure at the impact point and the counterpoint:the intracranial pressure at the impact site of deceleration injury has a large value and a large range of changes,and the intracranial pressure at the counter site has a greater range of changes and a longer duration.Conclusion:Virtual anatomy technology can fully obtain damage characteristics and damage distribution information,laying a foundation for analyzing and revealing the mechanism of craniocerebral injury.Deep learning technology has the advantages of objective and high efficiency in damage classification.Finite element simulation technology is used in the study of craniocerebral acceleration and deceleration It has obvious advantages in injury mechanism and stress transmission.The combination of three technical methods above have made up for the shortcomings of traditional techniques to a certain extent,perfected the traditional craniocerebral injury theory and empirical judgment,and can provide help for future forensic identification.