Researchers have created a simple interface that lets a person alter the look of a 3D model downloaded from the internet without changing its usefulness. It includes changing the model's colour, texture, and shape. 

An AI-driven tool simplifies the process of customizing 3D-printed models. Makers may quickly alter the appearance of 3D-printable products like assistive gadgets with Style2Fab without compromising the goods' functioning.

3D printer

As 3D printers have grown more affordable and widely available, a rapidly rising community of newbie makers is creating their products. Many amateur crafters accomplish this by accessing free, open-source archives of user-generated 3D models, which they download and print on their 3D printers.

However, adding bespoke design components to these models presents a significant problem for many producers because it necessitates using complex and expensive computer-aided design (CAD) tools, which is incredibly challenging if the original model representation is unavailable online. Furthermore, even if a user can add personalized aspects to an object, ensuring that those customizations do not interfere with the item's functionality necessitates a level of domain understanding that many inexperienced creators need to gain.

Personalized 3D-printable models

MIT researchers created a generative AI-driven tool that enables users to add unique design aspects to 3D models without sacrificing the functionality of the manufactured goods to aid producers in overcoming these difficulties. This application, dubbed Style2Fab, allows designers to customize 3D models of objects by describing the intended design using just natural language prompts. The user might then use a 3D printer to create the objects.

Style2Fab

Style2Fab uses deep-learning algorithms to divide the model into aesthetic and functional segments, speeding up design. It could expand medical manufacturing, empower novice designers, and make 3D printing more accessible. According to research, considering both the aesthetic and practical qualities of an assistive device increases the possibility that the patient will use it; nevertheless, physicians and patients may lack the expertise to personalize 3D-printable models.

A stylization tool would need to preserve the geometry of outwardly and internally useful segments, while nonfunctional, purely aesthetic elements might be customized. However, Style2Fab must identify the 3D model's functional components to do this. The method employs machine learning to assess model topology to track geometric changes like curves or angles where two planes join. Then, Style2Fab compares those segments to a dataset the researchers produced, which includes 294 models of 3D objects. A segment is designated functional if it closely resembles one of those parts.

Natural language prompt

After users agree to the segmentation, they provide a natural language prompt outlining their preferred design features. For instance, they may want a planter with a rough texture and several colours inspired by Chinoiserie aesthetics. Alternatively, they may demand a phone case that emulates the artistic style of Moroccan culture. The AI system, referred to as Text2Mesh, then endeavours to ascertain the visual representation of a three-dimensional model that aligns with the specifications provided by the user.

The Style2Fab software manipulates the model's aesthetic components, such as texture, colour, and shape adjustments, to achieve high resemblance. However, access to the functional segments is restricted. The researchers integrated the pieces into the backend of a user interface, which then performs automatic segmentation and stylization of a model depending on user inputs and a few clicks.

Conclusion

A study was undertaken including individuals with varying degrees of experience in 3D modelling, wherein it was observed that Style2Fab exhibited utility that varied depending on the competence of the individual engaging with it. Inexperienced users understood and used the interface to enrich designs. Moreover, it facilitated a conducive environment for engaging in exploratory endeavours, with minimal difficulty for initial participation.

Furthermore, Style2Fab has been found to enhance the efficiency of experienced users' processes. Moreover, using certain advanced features enhanced their precision in manipulating stylizations.

Image source: Unsplash