On our web site, users enter free form text like "Ribeye steak at Safeway for $8.99. Great deal in Palo Alto."
在我们的网站上，用户输入免费的表格文本，如“Safeway的Ribeye牛排，只需8.99美元。在Palo Alto很棒。”
Does anyone know of APIs that can take free form text and return an image that best represents the text?
有没有人知道可以采用*格式文本并返回最能代表文本的图像的API？
Right now, we're feeding the whole text to Google Image Search, but sometimes the extraneous words (e.g., at, for, $8.99) skew the query and induce weird results.
现在，我们正在将整个文本提供给Google图片搜索，但有时无关的单词（例如，at，for，8.99美元）会使查询偏斜并导致奇怪的结果。
1 个解决方案
#1
1
Google spends tens of millions of dollars a year trying to improve their answer to precisely this question; it's what Image Search is all about. If you are dealing with a relatively narrow sub-domain, your best bet is to build up your own image collection to match based on known keywords.
谷歌每年花费数千万美元试图改善他们对这个问题的答案;这就是图像搜索的全部意义所在。如果您正在处理相对较窄的子域，最好的办法是根据已知关键字建立自己的图像集以匹配。
If you still want to use Google, then you might try a hybrid approach of submitting only known keywords to them. I'd be worried about copyright and term-of-use issues, though.
如果您仍想使用Google，那么您可以尝试一种仅向其提交已知关键字的混合方法。不过，我会担心版权和使用条款问题。
#1
1
Google spends tens of millions of dollars a year trying to improve their answer to precisely this question; it's what Image Search is all about. If you are dealing with a relatively narrow sub-domain, your best bet is to build up your own image collection to match based on known keywords.
谷歌每年花费数千万美元试图改善他们对这个问题的答案;这就是图像搜索的全部意义所在。如果您正在处理相对较窄的子域，最好的办法是根据已知关键字建立自己的图像集以匹配。
If you still want to use Google, then you might try a hybrid approach of submitting only known keywords to them. I'd be worried about copyright and term-of-use issues, though.
如果您仍想使用Google，那么您可以尝试一种仅向其提交已知关键字的混合方法。不过，我会担心版权和使用条款问题。