Classification and Retrieval of Commodity Images Oriented to Internet Marketing

Commodity display images oriented to Internet marketing play an important role in the supply and demand interaction two-way communication between personnel consumers. Traditional commodity image classification is mainly conducted manually by staff of online shopping platforms or store maintenance staff, which heavy workload, high cost low efficiency. To this end, article studies retrieval marketing. In study, coarse-grained emotion taken as priori information, network based on joint polarity detection constructed. This discusses association rules color texture images, shape, styling features contained concrete commodities. Besides, puts forwards emotion-based method marketing, presents a train thought method. The experimental result verifies effectiveness images.