SBIR Phase I: Personalizing Online Clothing Shopping

这个小企业创新研究（SBIR）第一阶段项目的更广泛的影响/商业潜力将是改善服装在线购物的体验，目前是美国最大的在线细分市场，去年的销售额为510亿美元，是数百万消费者和数千家公司的焦点。 该项目的创新将提高发现和比较服装外观和风格的能力，允许在拥有数百万件商品的日益庞大的市场上进行更有效的购物，并提高在线服装市场的效率。 此外，该项目将提高计算算法自动解析和理解服装风格的能力，这是构建计算机以了解我们日常世界的一部分。这个小企业创新研究（SBIR）第一阶段项目将开发识别、比较和预测购物者对服装风格偏好的新技术。 这些将基于计算机视觉来识别服装款式的视觉特征，以及机器学习来从购物者在购物时的互动中建立购物者偏好模型。 发展将包括表征学习的服装风格的视觉外观，以及对风格的个人偏好的因素。 这些模型将被用于服装购物的自动搜索和推荐系统，为机器学习支持的个性化开辟新的商机。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project will be to improve the experience of online shopping for clothing, currently the largest segment online in the US with $51 Billion in sales for the last year and a focus for millions of consumers and thousands of companies. The innovation in this project would improve the ability to discover and compare the visual appearance and style of clothing items, allowing more effective shopping over an increasingly large marketplace with millions of items, and improving the efficiency of the online clothing market. In addition, the project will improve the ability of computational algorithms to automatically parse and understand clothing style, part of building computers to understand our daily world.This Small Business Innovation Research (SBIR) Phase I project will develop novel techniques for identifying, comparing, and predicting shopper preferences for clothing styles. These will be based on computer vision to recognize visual features of clothing styles and machine learning to build models of shopper preference from their interactions while shopping. Development will include representation learning for visual appearance of clothing style as well as for the factors that contribute to personal preference for style. These models will be used to automate search and recommendation systems for clothing shopping, opening up new business opportunities for machine-learning enabled personalization.