Electronic Science and Technology ›› 2021, Vol. 34 ›› Issue (10): 69-74.doi: 10.16180/j.cnki.issn1007-7820.2021.10.011

Previous Articles     Next Articles

Color Commodity Label Image Segmentation Method Based on SVM and Region Growth

JU Zhiyong,ZHAI Chunyu,ZHANG Wenxin   

  1. School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China
  • Received:2020-06-01 Online:2021-10-15 Published:2021-10-18
  • Supported by:
    National Natural Science Foundation of China(81101116)


The noise and gray scale growth in the traditional area may cause voids and over-segmentation, which will seriously affect the image segmentation effect. To solve this problem, this study proposes a color label image segmentation algorithm combining SVM and region growth. The algorithm uses the color information and texture of the label part and product part as positive and negative samples of the SVM, which effectively improves the efficiency and the accuracy of image segmentation. When extracting positive and negative samples, the improved region growth algorithm is used to segment feature areas and extract features. Theoretical analysis and experimental results show that the proposed algorithm has a mechanism for correcting error detection information, improves the segmentation efficiency, prevents the over-segmentation phenomenon of traditional regional growth algorithms during segmentation, and has good robustness, which lays the foundation for later expansion.

Key words: color image segmentation, region growth algorithm, support vector machine, morphological processing, image enhancement, labeling background, feature extraction, color space

CLC Number: 

  • TP751.1