Product search supports image categorization, which uses contextual information extracted from images to find images that exactly belong to the specified category. Product search is typically used for online shopping.
Generic search does not support image categorization. You do not need to specify a category ID when using generic search. Generic search is used for photo-stock company and photo-sharing websites.
Make sure all required fields are assigned a value and the body size of the request is less than 8 MB.
Generic search instances do not support reading data from OSS across regions. Make sure your instance and OSS are in the same region.
For Generic Search, you do not have to specify a category ID. This parameter has no effect.
For Product Search, you do have to specify a category ID when searching or inserting images. If no category ID is specified, the system automatically determines the category of the image. If a category ID is specified, the specified category ID is applied to the image.
If the crop parameter is set to true or not specified, Image Search crops the image and uses the cropped region as the effective region. If the region parameter is not specified, Image Search automatically determines the crop region.
If the crop parameter is set to false, Image Search does not crop the image. Feature extraction is performed based on the entire image area.
The SortExprValues value is a pair of two values, separated with a semicolon (;). The first value is a score that indicates the relevance. The higher the value, the more relevant the found image is to the search image. The relevance value falls in the range of [0, 7.33136443711219e+24]. When the result image is identical to the search image, the relevance value is the maximum value 7.33136443711219e+24. The second value is the Hamming distance between the search image and the result image in terms of their feature vectors. The higher the value, the less relevant the found image is to the search image. The sorting algorithm is based on the first value. Image Search does not provide an empirical threshold value for you to determine whether two images are generally relevant.
This issue may occur in the following situations:
- The specified OSS bucket and instance are in different regions. Currently, only the Singapore, China (Hong Kong), Japan (Tokyo), Australia (Sydney), and Germany (Frankfurt) regions are available.
- Your service role is not specified as Image Search Service.
- No authorization policy is created for your Image Search Service role, the authorization policy is incorrect, or you do not have permissions to access the data path in the specified bucket.
- The ARN is invalid, the bucket is invalid, the data path is incorrect, or the increment.meta file does not exist in the specified data path.