Multimodal fusion of stable isotope and infrared spectral data based on a GMM–ReliefF–BP framework: a novel approach for intelligent source attribution of plastic evidence in forensic science

Key Points

• Intelligent source attribution is achieved using multimodal fusion techniques for plastic evidence.
• The framework combines stable isotope and infrared spectral data to enhance accuracy in results.
• Utilizing a Gaussian Mixture Model–ReliefF–Backpropagation approach improves identification efficiency.
• This work may enable better forensic investigations by enhancing source attribution in plastic evidence.