Morphological Classification of Tobacco Ash Using a Deep Convolutional Neural Network: From Sherlock Holmes's "monograph" to modern computer vision
Keywords:
convolutional neural network, computer vision, forensic science, tobacco ash, classification, deep learning, trace evidenceAbstract
Arthur Conan Doyle's literary character Sherlock Holmes claimed the ability to visually distinguish the ashes of up to 140 varieties of tobacco. This paper examines how far that idea is realisable with modern artificial intelligence. We propose a methodological framework in which a convolutional neural network (CNN) is used to classify microscopic images of tobacco ash into five categories. The paper describes the data-collection protocol, model architecture, training procedure, and illustrative results. We show that the mineral and morphological features of ash: colour, texture, fibrous structure, carry sufficient discriminative information for machine classification, although forensic application requires combination with spectroscopic methods
Downloads
References
A. C. Doyle, "The Five Orange Pips," in The Adventures of Sherlock Holmes. Scotts Valley, CA: CreateSpace Independent Publishing Platform, 1891, 26 pp.
K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," in Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), 2016, pp. 770–778.
R. R. Selvaraju et al., "Grad-CAM: Visual explanations from deep networks via gradient-based localization," in Proc. IEEE Int. Conf. Computer Vision (ICCV), 2017.
M. Brady, "Artificial intelligence approaches to image understanding," AI Memo, Artificial Intelligence Laboratory, MIT, Cambridge, MA, pp. 205–264.
M. Albiez and M. Fischer, AI for Advanced Image Analysis. Oberkochen, Germany: ZEISS, 2023, 56 pp.
I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning. Cambridge, MA: MIT Press, 2016, 801 pp.
F. Chollet, Deep Learning with Python. Shelter Island, NY: Manning Publications, 2017, 350 pp.
