Texture Features Based Mouse Image Classification for Causes of Death Prediction
摘要
Forensic science is a discipline that is based on medicine and combines knowledge of law, biomedicine, bioinformatics, pharmacy, and related natural and social sciences. In recent years, the process of using texture features to solve problems in forensics has been developing, and this paper proposes to use texture features to classify mouse images, using Inception-V3 and ANN networks to classify datasets. Results show that the accuracy of using Inception-V3 network classification reaches 55.56%, which is greater than accuracy of ANN.