Image data set for AI-aided printed line smearing analysis of the roll-to-roll screen printing process for printed electronics

A total of 20 images were collected with cross-directional (CD) printed lines of various line widths and smearing areas using the in-house roll-to-roll screen printing system. Then these images were labeled pixel-wise into three classes: smearing, printed line, and background, and labels were saved as one channel 8-bit images with corresponding intensity 1, 2, and 3. This data set was used to train a U-Net-like deep convolutional neural network (DCNN) to detect continuous printed line smearing defects.

本研究采用自研卷对卷丝网印刷系统，采集了20幅包含不同线宽横向（Cross-directional, CD）印刷线及污渍区域的图像。随后对上述图像进行逐像素标注，划分为污渍、印刷线与背景三类，标注文件以单通道8位图像格式存储，各分类对应的像素强度分别为1、2、3。本数据集用于训练类U-Net深度卷积神经网络（Deep Convolutional Neural Network, DCNN），以实现连续印刷线污渍缺陷的检测。