Fabric image search can obtain technological parameters of existing similar fabric images to guide production, saving lots of labor and material resources in the proofing process. Traditional fabric search methods require manual assistance, being time-consuming and subjective. Studying automatic search method based on feature engineering can improve the precision and efficiency of fabric retrieval.
This paper presents a novel image search method for plaid fabrics based on handcrafted features. First, local texture descriptors are extracted by the local binary pattern on the separated images which are processed by Fourier transform. Key-point texture descriptors are extracted by Scale-Invariant Feature Transform (SIFT) and Vector of Locally Aggregated Descriptors (VLAD). Second, color moments with image partitioning are extracted to characterize spatial color information of plaid fabric images, and color coherence vector is adopted to assist in characterizing spatial color features. Third, the similarities of the three features are calculated and fused by the weight assignment to realize the plaid fabric image retrieval.
To verify the proposed strategy, 44,000 fabric samples are collected from the factory to build the image database as the benchmark. Experiments show that precision and recall at rank five reach to 74.4 and 52.6%, respectively, and mAP reaches to 0.718. Results prove that the proposed strategy is feasible and effective, which can realize plaid fabric image search fast and efficiently.
Experiments prove the feasibility, effectiveness and superiority of the proposed method compared to existing methods. The proposed method can help the fabric manufacturing factory save lots of labor and material resources in the process of fabric design, manufacturing and sale.
