Abstract:
Habesha women’s dress is one of the clothing cultures that women wear in most areas of
Ethiopia, and it is particularly common in the Amhara region. Although the embroidery
design of Habesha women’s dress appears to be the same, it has a different texture, shape,
and color. This embroidery design is represented by the Awi, BahirDarZege, Sekota,
SemenShewa, RayaKobo, Lalibela, Wollo, Gojjam, and Gondar Habesha women’s dress
cultures dressing styles. However, as fashion design on clothing develops, the Habesha
women’s dress pure cultural embroidered design is replaced with a different fashion style,
and the Habesha women’s dress embroidery design cultural identity is lost. There is a lot
of research on clothing patterns and color detection, recognition, and classification.
However, it's challenging to classify clothes with the same pattern, texture, and color into
various classes due to the problem of uncontrol environment, segmentation of ROI, color
space, and feature extraction of the embroidery design of Habesha women’s dress. The
goal of this study was to use an ensemble feature extraction approach to identify those
patterns.
A total of 8100 Habesha women’s dress embroidery image were collected from different
particular embroidery designer for different cultural clothe dress. We have used 900 images
for every nine classes. After getting the images, we applied different image pre-processing
techniques. Also, we have done feature extraction, by applying CNN, SGAN, and GLCM.
For classification, we have used SVM as a linear kernel. We have developed and trained
models such as CNN + GLCM (accuracy: 95.80%), SGAN + GLCM (accuracy: 97.84%),
CNN + SGAN (accuracy: 97.96%) and SGAN + GLCM + CNN (accuracy: 98.52%) with
PCA. All models have used linear kernel SVM classifiers. Finally, an ensemble of SGAN
+ GLCM + CNN with the SVM classifier achieved 98.52% accuracy. This combined
feature extractor performance has better accuracy than the individual for identifying
Habesha women’s dress embroidery design.
Keywords – CNN, Embroidery, GLCM, Habesha women’s dress, PCA, SGAN