Restorations
Restore lost details and decompress images
Last updated
Restore lost details and decompress images
Last updated
Image restorations are one of the most important operations to perform with Claid API. You can choose from a wide range of AI-powered neural networks to decompress
, upscale
, and polish
images for different processing workflows.
The upscale
operation allows you to improve the overall quality of an image by rendering out new pixels. The upscale operation works along with Resizing through numeric or percentage values for upscaling.
Upscaling is turned off by default. It can be defined by setting the upscale
object to one of the following upscaling methods below.
The lower the image resolution is the less information it has, making the upscaling more difficult. We recommend you to try different upscaling methods to increase the chances of finding consistent, quality results.
Name | Type | Description |
---|---|---|
| string | Used on small low quality product, real estate and food images. |
| string | Used on high-quality images and photos with barely readable text. |
| string | Used on drawings, illustrations, paintings, cartoons, anime, etc. |
| string | Used on images containing people |
| string | Used on photos of people, nature, architecture, etc. taken with phones or digital cameras. |
This code sample combines the use of restorations
and resizing
operations. For more details, see Resizing.
Tip: Click on the image to inspect it closer.
The decompress operation allows you to remove distortion and artifacts caused by image compression. It can be used in conjunction with the upscale
operation to prepare an image for upscaling.
Decompress is turned off by default. It can be configured by setting the decompress
object to either moderate
or strong
, depending on how heavily compressed images are. If you are not sure whether to apply a decompression – you can choose auto-select mode by setting the decompress
object to auto
.
Possible values:
Value | Type | Description |
---|---|---|
| string | Removes JPEG artifacts from the image. |
| string | Removes JPEG artifacts more aggressively than moderate. |
| string | Automatically detects and removes JPEG artifacts if needed. |
The polish operation allows you to redraw image parts making them sharper, yet keeping the original structure. It can be used individually, or in conjunction with the Upscale (we recommend Smart Enhance upscale) operation to make the image look even sharper and more realistic.
polish
operation currently has size limitations: the target image size must not exceed 16 MP (e.g., 4096 x 4096 pixels, 5000 x 3500 pixels). Our team is working on increasing max size.