Architecture for responding to a visual query

摘要: A visual query such as a photograph, screen shot, scanned image, video frame, or an image created by content authoring application is submitted to search system. The system processes the sending it plurality of parallel systems, each implementing distinct process. These systems may include but are not limited optical character recognition (OCR), facial recognition, product bar code object-or-object-category named entity and color recognition. Then at least one result sent client In some embodiments, when containing text element non-text element, includes for image-match element.