Skip to content

Cloudinary’s Latest Release: Scaling Image and Video Editing with AI-powered Automation

In this month’s blog you’ll find the latest product innovations from Cloudinary to improve your creative and development team’s productivity and ensure you’re delivering first-class visual experiences to your customers. 

Starting with some new AI-powered capabilities that help you perform mundane image and video editing tasks at scale. Based on popular demand, we have also released capabilities to manage related assets easily and enhance animated image optimization. 

(If you missed the December edition, you can find it here.)

Drop shadow is a commonly applied visual effect that gives the impression that the object in an image exists in a 3D space. The effect is used frequently in catalog display, advertising and traditional e-commerce scenarios. 

The new AI – Drop Shadow transformation automates the creation of a drop shadow effect for an object, or set of objects, in an image. 

One of the challenges with applying this effect is it often requires expensive creative tools and  technical expertise, especially when transforming  complex shapes and contours. Moreover, applying drop shadow to hundreds of images can be daunting even for enterprise-scale teams.  Now your team can add shadows using a single URL parameter (e_dropshadow). Specify the direction, angle, and spread of the light source for a perfect shadow of the object in the image. 

Here’s a cool example of applying AI – Drop Shadow in a typical e-commerce example:

Original
With Drop Shadow Effect

      

image/upload/e_dropshadow:azimuth_210;elevation_60;spread_10/v1675188603/bracelet-nobg.png

Another common image editing challenge is enlarging it without any loss in quality or resolution, often referred to as upscaling. For instance, brands that want to upscale a user-generated image across larger real estate on their website or mobile app often run into the issue wherein the image becomes grainy after upscaling because the original image is low resolution. 

Take the example of this image of Brooklyn Bridge. 

Original

Scaling this image to four times its dimension typically produces a grainier version but the new AI – Super Resolution Upscale transformation upscales the image without any loss in resolution. 

Use the slider in the center and slide to the extreme right to see the original image simply scaled up in size and then slide to the extreme left to see the original image upscaled using e_upscale transformation.

/image/upload/e_upscale/brooklyn_bridge.jpg

Before: 4x Larger, No Upscale
After:
Intelligently Upscaled Using e_upscale

Switching gears from AI innovations to asset management, introducing a new asset management capability – “Related Assets”.  

Related Assets feature creates a bi-directional connection between two media assets using a public ID. This offers greater flexibility in discovering, tracking and using disparate asset types such as audio transcript with a video asset. Another scenario where related assets feature could be valuable is when building “shop the look” product relationships to increase upsell or cross-sell opportunities on your e-commerce site. 

The Related Assets feature is currently available only via our RESTful Admin API. Support for SDKs is in the works.

Another key January innovation is out-of-box support for the Animated AVIF format within our popular optimization capability – f_auto. “Animated AVIF” is a modern image format that ensures the best size compression for an animated image without loss in visual quality. It can result in up to 90% byte savings compared to the most commonly used “animated gif” format. 


Animated AVIF is currently not supported on all browsers and that’s why it cannot be used in every scenario. With our latest release, whenever animated images are delivered using f_auto, you can now ensure that the animated AVIF format is automatically selected for browsers that support it so you get the best possible size compression when possible. Currently this feature is available by default to all our new customers. Existing customers wanting to enable this feature in their Cloudinary product environment will need to contact support.

Rounding out the January innovations is a proof-of-concept version of AI-powered video summarization capability by Cloudinary Labs. 

There is often a significant drop in engagement for videos after 2 minutes (source: Wistia). In such scenarios, a short video summary or preview can help convey the key message and maintain engagement. We currently support the ability to programmatically generate intelligent video previews with the e_preview transformation. 

Taking this concept further, we are introducing an interactive application for users to upload a video and automatically generate a brief summary of that video. Users can specify the desired length for the video summary too. This can be a particularly helpful tool for brands looking to create short marketing highlights, summary clips for use on social media and more! 

Built by Cloudinary Labs, this tool is now available in a proof-of-concept version that you can try for free. Take it for a  quick test drive and let us know what you think. We need your feedback!

That wraps up our key product highlights for the month of January. Stay tuned for more updates!
If you are new to Cloudinary and want to learn more, sign up for a free account.

Back to top

Featured Post