From Camera to Code: Vidu Seeks to "Reinvent" Photography with Image Compositing, Using 7 References to Create True-to-Life Images from Scratch

ShengShu Technology launches Vidu's new Reference-to-Image feature. Aiming to change the definition of photography, Vidu's new image compositing feature "generates" photographs from seven reference images and accurately preserves the image consistency of large text.

SINGAPORE, Sept. 8, 2025 /PRNewswire/ -- Vidu, the flagship product of ShengShu Technology and a pioneer in generative AI video, is launching Reference-to-Image as the startup aims to reinvent photography itself. Reference-to-Image is an industry breakthrough, enabling image editing or generating pictures with the industry's highest degree of consistency using seven reference images.

Until now, creating a 'photograph' or staging a shot with AI meant accepting whatever the model produced, with little ability to guide or refine it. Limited to just three reference images on most platforms, reference-to-image found its first users among brands and e-commerce companies. Vidu's Reference-to-Image feature is a historic turning point that redefines the perception of photography. Being able to support seven image references crosses the threshold where true-to-life photography can be generated with remarkable image consistency and high-fidelity realism as envisioned, without ever touching a camera and at a fraction of the cost of staging photo shoots.

Vidu Raises the Bar: More Images Result in Greater True-to-Intent Results

Vidu Reference-to-Image's support for up to seven reference images is a milestone technical leap, fueled by a marked improvement in the semantic understanding capability. Vidu now interprets the relationships between multiple images with far greater consistency, which until today was an area that was highly complex and error-prone even with only three reference inputs. Now, adding up to seven references produces outputs that are not only more realistic but also more faithful to the intended prompt, setting Vidu far apart from the competitors that often stray from user intent.

In fact, the precise orchestration of generating a photograph from scratch is now just a simple text prompt away, rendering photography virtually optional. For instance, last minute post-production edits for filmmakers or photographers without the expensive costs of models and props for a reshoot, or a staged outdoor wedding shoot ruined by rain, can now be salvaged with reference images.

This could include a new style of the bride's dress, a shot of both the groom and the bride, a bouquet of flowers, a sunny backdrop of the same location, a standing photo of a missing family member that was out sick, and two props. With these seven references and a text prompt, Vidu's model can blend them seamlessly, remove distractions like tourists, and recreate a believable sunny wedding photo, even if the real moment was spoiled by weather and none of the footage was usable.

Vidu Reference-to-Image Excels in Image Consistency

Vidu, even when compared to leading generative AI models, offers unmatched image and character consistency, along with natural image blending for richer and more realistic details. In fact, among the industry's leading competitors, while Vidu leads for its reference-to-image feature, it stands head-to-head or even outcompetes in overall performance.

Diving deeper into its reference-to-image capability, reference visuals and text are carried over from the original source image with a higher degree of clarity than all other generative platforms. For example, when inputting several images, including a rice cooker with a text logo, a female model, and a bowl of rice, Vidu's ability to retain realism enables the rice cooker and a bowl of rice to be juxtaposed together in the image in a believable way. Vidu's output results in a model elegantly holding a bowl of rice in her left hand, while hunched over with her right elbow leaning over the rice cooker, while the logo (without distortion) is clearly legible.

In fact, image consistency not only enables Vidu to generate precise photographs and designs, but it also enables its users to edit photos more accurately, while still preserving clear, readable text. This may include tweaking a candid selfie that doesn't look right; A/B testing an interior design concept for a client; swapping out products within an e-commerce advertising shoot; or adding an extra character last minute into the final draft of an animated series.

For more precise edits, Vidu has been significantly improved to offer instant image editing capabilities rivaling leading editing suites. With one image and a prompt, you can replace or add objects to an image, which dramatically reduces editing time and costs. Some of these editing features launching today include:

• Remixing enables multiple input images to be freely composited into a single image that preserves high consistency with the reference images, but also an overall visual plausibility.
• Partial Replacement can change the look and feel of an object, like the color of an outfit or umbrella.
• Object Replacement enables users to change or replace objects within the image entirely, which could result in a book that the subject of the photo is holding being replaced by a traditional point-and-shoot camera.
• Adding Objects introduces the ability to believably embed new objects within the image, without disrupting the context and consistency of the photograph.

Photography is on the brink of a seismic shift, as generative AI begins to rival creative control and fidelity, once exclusive to cameras and lenses. This update from Vidu isn't just about replacing tools. From an industry-leading defining first-to-market generative video features like subject consistency (even before Sora), Vidu is redefining what it means to capture, compose, and even imagine an image.

To learn more about Vidu visit https://www.vidu.com

Vidu API: platform.vidu.com 

About ShengShu Technology

Founded in March 2023, ShengShu Technology is a world-leading artificial intelligence company, specializing in the development of Multimodal Large Language Models. Driven by innovation, the company delivers cutting-edge MaaS and SaaS products that revolutionize creative production by enabling smarter, faster, and more scalable content creation. With its flagship video generation platform Vidu, ShengShu Technology's solutions have reached more than 200 countries and regions around the world, spanning fields including interactive entertainment, advertising, film, animation, cultural tourism, and more.

SOURCE ShengShu Technology