Scenario 1

Scenario 1: Product-specific Sentiment Analysis

This use-case scenario focuses on product-specific identification within images and sentiment analysis based on textual information and reviews of the identified product. The steps of this scenario are the following:

1. Web scraping – collect product’s images, descriptions and reviews from related websites
2. Image and product name classification – train the system on product’s image by providing different images of the same product. This done also by text classification, in order to find as many images as possible for the same product.
3. Sentiment analysis and summarization – the textual scraped data will be used for sentiment analysis and auto summarization of the product description
4. Image query – when an image is uploaded to the system, the system searches for the relevant product and displays it sentiment and short review of the product.