We found a match
Your institution may have rights to this item. Sign in to continue.
- Title
Age Determination of Chrysomya megacephala Pupae through Reflectance and Machine Learning Analysis.
- Authors
Zhang, Xiangyan; Qu, Hongke; Zhou, Ziqi; Chen, Sile; Ngando, Fernand Jocelin; Yang, Fengqin; Xiao, Jiao; Guo, Yadong; Cai, Jifeng; Zhang, Changquan
- Abstract
Simple Summary: The duration of pupa development on a cadaver holds the potential for estimating the time of colonization (TOC), which is often correlated with the postmortem interval (PMI) of decomposed bodies. Establishing an objective, precise, and efficient method for inferring pupa age has become paramount in forensic entomology. In this study, we observed temporal variations in the reflection spectrum of Chrysomya megacephala (Diptera: Calliphoridae) pupa, detectable through hyperspectral imaging (HSI). Additionally, we proposed that the eXtreme Gradient Boosting Regression (XGBR) model represents an optimal approach for estimating pupa development time based on HSI data. Estimating the age of pupa during the development time of the blow fly Chrysomya megacephala (Diptera: Calliphoridae) is of forensic significance as it assists in determining the time of colonization (TOC), which could help to determine the postmortem interval (PMI). However, establishing an objective, accurate, and efficient method for pupa age inference is still a leading matter of concern among forensic entomologists. In this study, we utilized hyperspectral imaging (HSI) technology to analyze the reflectance changes of pupa development under different temperatures (15 °C, 20 °C, 25 °C, and 30 °C). The spectrograms showed a downtrend under all temperatures. We used PCA to reduce the dimensionality of the spectral data, and then machine learning models (RF, SVR-RBF, SVR-POLY, XGBR, and Lasso) were built. RF, SVR with RBF kernel, and XGBR could show promise in accurate developmental time estimation using accumulated degree days. Among these, the XGBR model consistently exhibited the most minor errors, ranging between 3.9156 and 7.3951 (MAE). This study has identified the value of further refinement of HSI in forensic applications involving entomological specimens, and identified the considerable potential of HSI in forensic practice.
- Subjects
PUPAE; MACHINE learning; TIME perception; FORENSIC entomology; REFLECTANCE; BLOWFLIES; FORENSIC genetics
- Publication
Insects (2075-4450), 2024, Vol 15, Issue 3, p184
- ISSN
2075-4450
- Publication type
Article
- DOI
10.3390/insects15030184