Abstract In this paper, a novel classification method of assessing garment sewing stitch based on amended bi-dimensional empirical mode decomposition (ABEMD) has been introduced. Two parameters that characterise garment sewing stitch, average area and standard deviation, have been defined based on the grey value of pixels. Experimental results showed that when the window size is 512×128 pixels with regard to average area, the threshold can be decided as 6.00, 5.50, 5.30 and 4.00 for five different grades , respectively. Meanwhile, with regard to standard deviation, the threshold can be decided as 48.00, 40.00, 30.00 and 20.00, respectively. It is demonstrated that the parameters are effective in discriminating sewing stitch images in terms of the grades when used as inputs for the ABEMD. The performance of the algorithm on different garment status is significantly reliable.
Autex Research Journal – de Gruyter
Published: Dec 31, 2013