الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 32 |  |  |
| | | from | 32 | | |
Abstract
Defect detection is accomplished by human experts, this imposes to limited accuracy,consistency and efficiency, also uncertain and biased inspection results are produced.Elimination of sewing defects is one of the importance requirements to increase the quality of the garments. The detection of the sewing defects should be as quick as possible to decrease the consuming time to inspect these defects.Therefore, the aim of this work is to detect the features of some common stitch defects; (Nonendless of the seam line, skipped stitch or floated stitch, small hole, big hole, missed stitch and seam pucker) by analyzing them using three different methods analysis; they are Wavelet Transform, Image Features and Descriptive Features, and classify them by using the Artificial Neural Networks.The results were divided into two sections; section one to get the high recognition rate of the three procedures and the second one to test the Reality of the Neural Networks of the three procedures. To obtain high recognition rate, the size of the images was decreased to be (40* 35) pixels and the hole defect was divided into two categories according to its size. The Image Features procedure achieved the highest Reality rate in the three procedures.