This document mainly introduces some common problems you may encounter when you perform the BOS data processing and the appropriate handling methods.

Usage Limitations

What formats of images are supported for image processing?

• The BOS image processing supports the input and output in the mainstream formats. The supported input formats of original images include jpg, png, bmp, webp, heic, and gif. The supported output formats of destination images include jpg, png, bmp, webp, heic, and gif. BOS provides the format conversion, i.e., the output format of image is different from the input format of the original image. heic and webp are advanced compression formats, and a specific fee is charged according to the number of calls. For advanced compression pricing, see Introduction to BOS Billing Items and [BOS Price](https://cloud.baidu.com/doc/BOS/s/Tkev2iiac#Image Processing Cost).

Does the image processing have any limitations on the input of the original image?

• BOS supports the input image with a size of no more than 20 MB and a width and height of no more than 4,096 pixels. It supports the output image of the processing result with the width and height of no more than 4,096 pixels, including the intermediate output image between pipeline and subcommand.

How to Use

How can image processing be called?

• BOS supports the calling and reading of original images through THE API directly and can also carry the processing command during reading and writing. If it carries the processing command during file reading, it reads the image processed. If it carries the processing command during file writing, it writes a new processed image file, and the original image file is still saved. Up to 4 processing operations can be specified in the same command, i.e., 4 actions. The same parameter cannot be repeated in an action. To facilitate the use, you can also encapsulate an action as a style for calling in advance. For detailed rules, please see Usage Rules for Image Processing.

How to create an image processing style?

• You can add or delete the style through the BOS console. A single bucket supports the creation of up to 400 image processing styles. For detailed operation steps, please see [Add/Delete Image Styles](https://cloud.baidu.com/doc/BOS/s/Qkbis7yxm#Add/Delete Image Styles).

What processing content does the image processing support?

• The BOS image processing has several processing modes, including basic image processing, intelligent image processing, image review, and OCR recognition capabilities as a whole. Howevver, the usage modes of different capabilities are different. The basic image processing and intelligent image processing are used through the native Image Processing Rules of BOS. The image review provides services through the separate Interface and API. Also, it cannot be used through the native image processing rules of BOS. If you want to use the OCR feature, such as universal character recognition or universal object recognition, you can trigger other Baidu AI Cloud-related products through the Event Notification Feature.

Product Price

Is the image processing charged?

• The BOS basic image processing, intelligent image processing, and advanced compression are paid services, which are charged according to the called image size (GB) or the number of calls. For image review and other applications called through the event notification function, BOS only charges API reading and writing fees. For the billing of image processing, please see Introduction to BOS Billing Items and [BOS Price](https://cloud.baidu.com/doc/BOS/s/Tkev2iiac#Image Processing Cost).