Abstract：Determining line-crossing sequences between stamp inks and laser printing toner forms a core component of questioned document examination, offering critical evidence for authenticating questioned documents. Current methodologies are hindered by examiner dependency and limited applicability, failing to meet the demands for non-destructive, automated, and high-precision detection. While spectrochemical techniques combined with pattern recognition show promise in non-destructive analysis, they remain insufficient for investigating ink-toner interpenetration cases and model accuracy. This study proposes a novel approach that integrates visible-near-infrared hyperspectral imaging (Vis-NIR HSI) with machine learning, utilizing 10 000-pixel spectra per sample from three ink types and laser printers. Initially, a systematic multi-angle spectral difference calculation strategy was employed to quantify the temporal characteristics of material deposition sequences accurately. Following this, a new method that combines standard normal variate (SNV) spectral preprocessing with the interval variable iterative space shrinkage algorithm (iVISSA) was used to enhance model performance by minimizing noise and intelligently selecting wavelengths. In conclusion, a comprehensive comparative analysis of six machine learning models, including logistic regression (LR), support vector machines (SVM), and random forests (RF), was conducted to develop a high-accuracy classification system for questioned document examination. Experimental results demonstrate that the log ratio method effectively improves spectral differentiation. The combined application of SNV preprocessing and iVISSA-based characteristic wavelength selection successfully reduces data redundancy while significantly boosting model performance. Among the evaluated algorithms, SVM, XGBoost, and LightGBM emerged as the top models, showing robust capability in determining line-crossing sequences between compatible stamp inks and laser printing toners. Validation confirms the method's superior generalization ability for questioned document examination. This study addresses limitations by demonstrating the feasibility of integrating Vis-NIR HSI with pattern recognition to analyze line-crossing sequences of compatible stamp inks. This approach enhances conventional techniques, adding substantial judicial value in fighting document fraud.
李昌盛，高树辉，李开开. 基于高光谱技术的朱墨时序光谱差异特征分析及分类[J]. 光谱学与光谱分析, 2025, 45(10): 2804-2815.
LI Chang-sheng, GAO Shu-hui, LI Kai-kai. Feature Analysis and Classification of the Line-Crossing Sequences
Between Stamp Inks and Laser Printing Toner Based on
Hyperspectral Techniques. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2025, 45(10): 2804-2815.
[1] Li B. Journal of Forensic Sciences, 2016, 61(3): 809.
[2] Ministry of Public Security, the People's Republic of China(中华人民共和国公安部). GA/T 1958—2021 Forensic Sciences—Specifications for Examination of the Sequence Between Stamps and Strokes(GA/T 1958—2021 法庭科学朱墨时序检验规范), 2021.
[3] TAO Ke-ming, YAN Yu-hua(陶克明, 阎育华). Forensic Science and Technology(刑事技术), 1989, (4): 13.
[4] Lee K Y, Lee J, Kong S G, et al. Forensic Science International, 2014, 234: 120.
[5] OUYANG Guo-liang(欧阳国亮). Journal of Henan Judicial Police Vocational College(河南司法警官职业学院学报), 2020, 18(2): 102.
[6] LI Biao, FENG Ming-shuai(李 彪, 冯明帅). Guangdong Gongan Keji(广东公安科技), 2012, 20(1): 10.
[7] SUN Hua-qing, DONG Jun, XU Yan-yan, et al(孙华清, 董 军, 许燕燕, 等). China Public Security(Academy Edition)[中国公共安全(学术版)], 2012, (1): 113.
[8] XIE Peng, LI Biao(谢 朋, 李 彪). Journal of People's Public Security University of China(Science and Technology) [中国人民公安大学学报(自然科学版)], 2003, (2): 36.
[9] Wu X, Fang F, Li B. Journal of Forensic Sciences, 2019, 64(6): 1761.
[10] Wu X, Li B, Ouyang G. Journal of Forensic Sciences, 2021, 66(4): 1545.
[11] Kaur R, Kaur G. International Journal of Medical Toxicology and Legal Medicine, 2021, 24(3 and 4): 73.
[12] Meneghetti M, Mancini L, Caligiore G, et al. Forensic Science International, 2024, 360: 112051.
[13] Bojko K, Roux C, Reedy B J. Journal of Forensic Sciences, 2008, 53(6): 1458.
[14] Morais D R de, Campêlo J de M, Razzo D, et al. Analytical Methods, 2020, 12(7): 951.
[15] Barac M, Filko A, Siketic Z, et al. Forensic Science International, 2022, 331: 111136.
[16] Mathayan V, Sortica M, Primetzhofer D. Journal of Forensic Sciences, 2021, 66(4): 1401.
[17] Liu S, Yang Y, Zhang Y, et al. Journal of Forensic Sciences, 2024, 69(6): 2148.
[18] Dirwono W, Park J S, Agustin-Camacho M R, et al. Forensic Science International, 2010, 199(1-3): 6.
[19] Wang Y, Li B. Science & Justice, 2012, 52(2): 112.
[20] Silva C S, Pimentel M F, Honorato R S, et al. Analyst, 2014, 139(20): 5176.
[21] Borba F de S L, Jawhari T, Honorato R S, et al. Analyst, 2017, 142(7): 1106.
[22] Brito L R, Braz A, Honorato R S, et al. Forensic Science International, 2019, 298: 169.
[23] Brito L R, Chaves A B, Braz A, et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2019, 223: 117287.
[24] Freden S C, Mercanti E P, Becker M A. Third Earth Resources Technology Satellite-1 Symposium: Section a-B. Technical Presentations. Scientific and Technical Information Office, National Aeronautics and Space Administration, 1974.
[25] Chavez Jr P S. Remote Sensing of Environment, 1988, 24(3): 459.
[26] Barnes R J, Dhanoa M S, Lister S J. Applied Spectroscopy, 1989, 43(5): 772.
[27] Deng B-C, Yun Y-H, Ma P, et al. Analyst, 2015, 140(6): 1876.
[28] Li W, Chen C, Su H, et al. IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(7): 3681.
[29] Chaurasia V, Pal S. SN Computer Science, 2020, 1(5): 270.
[30] Herath H C M. Performance Evaluation of Machine Learning Classifiers for Hyperspectral Images. 2021 IEEE 21st International Conference on Communication Technology (ICCT). 2021, 1216.
[31] Vanneschi L, Silva S. Decision Tree Learning in: Lectures on Intelligent Systems. Cham: Springer International Publishing, 2023, 149.
[32] Jafarzadeh H, Mahdianpari M, Gill E, et al. Remote Sensing, 2021, 13(21): 4405.
[33] Samat A, Li E, Wang W, et al. Remote Sensing, 2020, 12(12): 1973.
[34] Zheng R, Jia Y, Ullagaddi C, et al. Food Chemistry, 2024, 456: 140062.
[35] XIE Shuo-yue(谢硕玥). Journal of Sichuan Police College(四川警察学院学报), 2018, 30(5): 14.
[36] Bhat S, Mansoor A, Georgescu B, et al. Scientific Reports, 2023, 13(1): 21097.
