Search

Herein, the challenge of accurately classifying the manufacturing origin of printing paper, including continent, country, and specific product, was addressed. One-dimensional convolutional neural network (1D CNN) models trained on infrared (IR) spectrum data acquired from printing paper samples were used for the task. The preprocessing of the IR spectra through a second-derivative transformation and the restriction of the spectral range to 1800 to 1200 cm-1 improved the classification performance of the model. The outcomes were highly promising. Models trained on second-derivative IR spectra in the 1800 to 1200-cm-1 range exhibited perfect classification for the manufacturing continent and country, with an impressive F1 score of 0.980 for product classification. Notably, the developed 1D CNN model outperformed traditional machine learning classifiers, such as support vector machines and feed-forward neural networks. In addition, the application of data point attribution enhanced the transparency of the decision-making process of the model, offering insights into the spectral patterns that affect classification. This study makes a considerable contribution to printing paper classification, with potential implications for accurate origin identification in various fields.

Effective rejection of time-varying clutter originating from slowly moving vessels and surrounding tissues is important for depicting hemodynamics in ultrasound color Doppler imaging (CDI). In this paper, a new adaptive clutter rejection method based on spectral analysis (ACR-SA) is presented for suppressing nonstationary clutter. In ACR-SA, tissue and flow characteristics are analyzed by singular value decomposition and tissue acceleration of backscattered Doppler signals to determine an appropriate clutter filter from a set of clutter filters. To evaluate the ACR-SA method, 20 frames of complex baseband data were acquired by a commercial ultrasound system equipped with a research package (Accuvix V10, Samsung Medison, Seoul, Korea) using a 3.5-MHz convex array probe by introducing tissue movements to the flow phantom (Gammex 1425 A LE, Gammex, Middleton, WI, USA). In addition, 20 frames of in vivo abdominal data from five volunteers were captured. From the phantom experiment, the ACR-SA method provided 2.43 dB (p <; 0.001) and 1.09 dB ( ) improvements in flow signal-to-clutter ratio (SCR) compared to static (STA) and down-mixing (ACR-DM) methods. Similarly, it showed smaller values in fractional residual clutter area (FRCA) compared to the STA and ACR-DM methods (i.e., 2.3% versus 5.4% and 3.7%, respectively, ). The consistent improvements in SCR from the proposed ACR-SA method were obtained with the in vivo abdominal data (i.e., 4.97 dB and 3.39 dB over STA and ACR-DM, respectively). The ACR-SA method showed less than 1% FRCA values for all in vivo abdominal data. These results indicate that the proposed ACR-SA method can improve image quality in CDI by providing enhanced rejection of nonstationary clutter.