| 摘要 |
An embodiment of the invention includes a network-accessible compute node, which includes a local storage storing reference images. Each reference image can depict one or more preferences, which can include a quality, a feature, a characteristic, an attribute, a type, and/or a form. Each preference can be associated with a distinctive pattern and a preference criterion. An embodiment includes an optimization module. The optimization module can learn the distinctive patterns from the reference images. The optimization module can also access a remote storage storing images of commodities and use pattern recognition to identify, from the remote storage, one or more images of commodities meeting the preference criterion selected by a user. Other embodiments are described and claimed. |
