Arvato Systems Revolutionizes 3D Visualization with Google Cloud
Imagejet boosts E-commerce VR
In the era of e-commerce and advanced digital business strategies, there is an increasing demand for photorealistic product images, which significantly determine commercial success. Current statistics highlight that visual representations of products are a central decision criterion for 87% of consumers during online shopping, while 67% emphasize image quality as the decisive factor for their purchasing decision. Studies suggest that image quality is even preferred over detailed product information (63%) and comprehensive descriptions (54%).
Imagine previewing a new couch in your living room before purchase through a detailed 3D representation, including realistic textures, shadows, and accessories tailored to your preferences. Such 3D asset creations are not only excellent for image production but also significant for immersive technologies like Augmented Reality (AR) and Virtual Reality (VR).
As a Google Cloud Premier Partner, we have introduced ‘imagejet,’ a groundbreaking, cloud-native solution for 3D image rendering. Thanks to the integration of powerful G2 VMs powered by NVIDIA L4 Tensor Core GPUs, imagejet offers businesses a straightforward platform for generating studio-quality images and GenAI training, enabling high-quality image production on a large scale.
Reimagine the Ordinary: Breaking New Ground in Digital Presentations
In the dynamic post-COVID retail landscape, the rising requirement for high-quality product visualizations is evident. Retailers, consumer goods manufacturers, and medium-sized enterprises see it as an occasion to rethink their digital strategies and seek efficient means to enhance their online product presence.
Traditional photography proves to be suboptimal: cost-intensive and rigid, it lacks flexibility and scalability. A professional shoot in a physical studio entails high investments for equipment and lighting, along with further limitations, such as the restricted representation of product variants.
Conversely, the challenging task of independently implementing a performant 3D rendering system arises. The development costs and infrastructure effort – including the acquisition of high-end NVIDIA GPUs and expensive licenses for 3D render software – create significant financial and technical barriers.
An independent implementation of an image production pipeline resembles a mammoth project that requires extensive, often unrelated expertise, in addition to financial expenditures. Imagejet offers an ideal solution here, as it is designed for regular production cycles and allows perfect reproducibility while maintaining high flexibility in the face of new requirements.
Optimizing 3D Visualization Processes with imagejet on Google Cloud
With our tool imagejet tool, we have significantly increased efficiency in 3D image production. The main concern of this development was to minimize the barrier to entry into 3D rendering processes, thereby decisively reducing costs and complexity for customers.
The core functionalities of imagejet include:
Simplified generation of images from 3D models, where previous photographic templates remain untouched.
A collaborative exchange space that optimizes the interaction between end customers and 3D agencies.
A seamlessly integrated system for monitoring and approving workflows and work processes.
Highly scalable image production capacities and efficient output of image material, supported by Google Cloud infrastructure and applying leading technologies like Pixar’s Universal Scene Description (USD) and standardized PIM data.
Multi-channel publication competence, enabling flexible distribution of a single asset across various sales channels, including AR and VR.
Gen-AI training data can be generated with imagejet.
For example, if a customer needs four million images for their product catalog, imagejet’s innovative approach, supported by the powerful NVIDIA L4 GPUs of Google Cloud and the use of optimized spot instances, could deliver this volume within just two weeks. A conventional approach using internal or locally operated 3D production pipelines would currently take up to two years.
Innovation Partnership with Google Cloud
In developing imagejet, we aimed to find a cloud provider that offers an extensive portfolio of cutting-edge technologies and solid partnerships in the tech industry. The key cooperation between Google Cloud and NVIDIA ensures that Google Cloud customers have access to forward-looking GPUs and robust availabilities.
Value-added services from Google Cloud, including Google Kubernetes Engine, Pub/Sub, Cloud Storage, Cloud Memorystore for Redis, and BigQuery, played a crucial role in enhancing imagejet’s performance.
The intensive collaboration with the Google Cloud Customer Engineering Team revealed that NVIDIA’s L4 Tensor Core GPUs are ideally suited for imagejet’s specific 3D rendering needs, which require precise 32-bit floating-point operations. With the latest NVIDIA L4 GPUs, Arvato Systems achieved a 160% performance increase in 3D rendering and reduced costs by 75% compared to the previously used NVIDIA A100 Tensor Core GPUs.
Imagejet represents a turning point in product visualization and offers an innovative method to produce high-quality 3D visuals more efficiently and cost-effectively. In a time when immersive and interactive shopping is becoming increasingly popular, imagejet promises to transform industrial practices sustainably.
Christian Scholz, VP of Cloud and Business Transformation at Arvato Systems, comments: “With support from Google Cloud, our solution imagejet will revolutionize the workflows of creatives, agencies, marketing teams, gamers, developers, and many more. We provide our customers with powerful tools that enable them to create synergies and craft digital realities.”
This blog article is based on the blog post by Google Cloud "How Arvato Systems makes 3D picture production easier, faster and cheaper with Google Cloud."
Written by
