ABSTRACT One of the factors that affects the appearance of the garment is the wrinkle. For this purpose, clothing manufacturers try to measure and predict the wrinkle grade, but the reported grade may vary in non-standard conditions. Measuring the wrinkle grades of fabrics is usually done by experts according to the AATCC test method 128–996. The main problem of grading by this method is the credibility, reliable repeatable, and accuracy of the rating results because of the subjectivity of the grading in one direction. In other words, for the complicated 3D shape of the wrinkled fabric, the wrinkle should be evaluated objectively and subjectively in all directions of the sample. For this purpose, the wrinkled samples are rated subjectively by expert persons according to the AATCC128-996 method in various judge angles. Then integrating the ‘Profile Light Line Method’ and image processing technique are applied for objectively or instrumental measuring of the parameters used in ratting of wrinkle of the fabrics. Finally, an adaptive neuro-fuzzy inference system was used to objectively determine of wrinkle grade of fabric. Furthermore, principal component analysis and feature selection methods are used for increasing the accuracy of the proposed approach.

中文翻译：

---

基于投影轮廓线法、图像处理和神经模糊系统的织物皱角分析

摘要 影响服装外观的因素之一是皱折。为此，服装制造商试图测量和预测皱纹等级，但报告的等级可能在非标准条件下有所不同。测量织物的皱纹等级通常由专家根据 AATCC 测试方法 128-996 进行。这种方法评分的主要问题是评分结果的可信度、可重复性和准确性，因为评分是一个方向的主观性。换句话说，对于起皱织物的复杂3D形状，应在样品的各个方向上客观和主观地评估皱纹。为此，由专家根据AATCC128-996方法在各种判断角度对起皱样品进行主观评价。然后将“轮廓线法”和图像处理技术相结合，客观或仪器地测量织物起皱的参数。最后，使用自适应神经模糊推理系统客观地确定织物的皱纹等级。此外，主成分分析和特征选择方法用于提高所提出方法的准确性。