This page is optimized for AI. For the human-readable: AI ENHANCED 3D MODEL MAKER

AI ENHANCED 3D MODEL MAKER

Project Idea Metadata

Project Idea Description

– What problem would you like to solve?​

Modelling a garment in 3D requires some hours work of an experienced model maker, some features do require the refined experience of a professional, others are simpler to solve and here is where we would like to use the support of AI.

The support of a software sewing the garments and uploading them to 3d clothing software would allow the professional to concentrate on qualitative portion of the process where the expertise is a game changer.

A larger and quicker availability of 3D blank garments would also empower the design team and allow them to be able to start working on new collections in 3D, even for fashion companies like Guess where models are many and varied.

– How does the proposed solution solve the problem?

Our solution proposes an innovative 3D modeling software enhanced by artificial intelligence, which revolutionizes the process of creating virtual garments:

– Input and Process: The software receives a pattern in .dxf format as input, representing the various parts of the garment in 2D. Using advanced AI algorithms, the system interprets the components of the pattern, understanding their layout and

the implicit sewing instructions.

– AI Processing and Virtual Assembly: The artificial intelligence analyzes the pattern and simulates the sewing process. The software virtually "assembles" the parts, replicating basic sewing techniques. Initially, to optimize resources and ensure a solid foundation, the system will focus on simple garments such as t-shirts. This process includes basic operations like joining side seams, attaching sleeves, and creating hems. As the project evolves, more complex garment structures and sewing techniques will be incorporated, gradually expanding the system's capabilities to handle a wider range of clothing items.

– Output and Compatibility: The result is a complete and realistic 3D model of the garment, maintaining the proportions and characteristics of the original design. This 3D model is generated in an agnostic format compatible with the major software platforms currently used (es. CLO3D), ready for further modifications and refinements. The project also explores the development of a prototype plugin that aims to facilitate a basic level of integration with this software platform. This preliminary integration would allow for an initial, limited transition between our semi-automatic generation process and manual refinement tools, serving as a proof of concept for future, more comprehensive integrations.

This innovative approach automates the most laborious and repetitive part of the 3D modeling process, allowing professionals to focus on the creative and high-quality aspects of clothing design.

– Who are the customers / users and what are the benefits?

The main users would be pattern makers that could move their focus from quantity to quality and will also acquire the competencies required by training an AI. Designers will benefit of it as the quantity of already sewn garment is bound to grow and as model makers will have more time to perfection the models they could work hand-in-hand on the new collection directly in 3D. As of today, this is not possible due to the short timing allowed by production calendar but moving towards a qualitative work would open up chances.

Being able to fulfill the whole flow of collection from design to sales campaign will also reduce the quantity of samples necessary.

- How does your challenge and proposed solution address the needs of Fashion & Lifestyle industry (incl. research, society and end users)?

Being able to use an efficient and quick 3D from the beginning of the creation flow ending in sales campaign and marketing, would result in more sustainability on the chain of value as less or any prototypes will be needed.

Same goes for samples, having the whole collection in 3D will allow sales on a digital showroom or even on the metaverse, cutting the necessity of people to travel to see the collection.

As of today, 3D in fashion is for early adopters, lots of companies, especially those with a high content of fashion, are still in doubt on how to deploy it.

Such a solution would pave the way for an innovative approach shred across the industry. Making it simple will help the technology to spread.

Companies will also ensure the retention of their employees allowing them to grow and use new technological tools to approach their jobs, this does not apply only to new generation but also to accompany the skill up of more experienced resources that may feel cut out of the technological process.


Implementation and risks:

– What is your Innovation Idea workplan and milestones?



As said, 3D in clothing is still an early bird topic, so we aim to create something from scratch in a not-yet mature technological environment. Therefore IB funds will be used for a feasibility study in three phases that will develop over a time span of three/four months

PHASE 1: Requirements Analysis and Knowledge Acquisition Milestone: Comprehensive project blueprint and successful knowledge integration:

– Formalize detailed project requirements

– Evaluation of the possible legal and patent limits in which to move

– Facilitate knowledge transfer from industry experts to research team

– Analyse current industry workflows and pain points

– Define key performance indicators (KPIs) for project success


PHASE 2: AI Model Research and Selection Milestone: Optimal AI model identified for virtual garment assembly:

– Research state-of-the-art AI models for 3D garment construction

– Evaluate models that emphasize Huma-AI Collaboration (HITL)

– Evaluate models based on accuracy, efficiency, and scalability

– Select and adapt the most suitable AI model for the sewing process

– Develop initial prototypes to validate model selection


PHASE 3: Dataset Development and Preprocessing Milestone: Robust and diverse dataset ready for AI training:

– Collect and curate diverse garment patterns

– Design and develop a universal 3D format for garment representation, ensuring interoperability with proprietary formats (e.g., CLO3D, Browzwear, Style3D)

– Implement data augmentation techniques

– Ensure dataset quality and representation of various garment types

Of course we are aware that this is only the introduction, but it’s absolutely necessary to make sure the project can land. Therefore, once the above portion is successfully completed, we already planned a timeline on actual development of the software, please refer to point 11 for details.

– What are the risks?

– Integration Limitations with Proprietary Software Risk:

Integration with CLO3D must be verified due to proprietary format, typical of all the clothing 3D software.

– Legal Challenges in Plugin Development Risk:

Cooperation with CLO 3D has to be put in place to make sure the plug in will be effective

– Limited and Inaccessible Datasets Risk:

Dataset limitations in terms of size and accessibility could hinder AI model training and system effectiveness so the workload to enhance it will be a huge effort from Guess side



Resources:

– What is your estimation considering needed resources and work packages?

– How are you planning to spend the innovation booster awarded amount in case you will win the application?

All received funds will be devoted to the research, any information, assistance or data needed from Guess is planned to be provided as additional part of daily activities.

– Feasibility Study: 3’000 CHF

· Collaboration with designers and experts for feasibility assessment.

· Interoperability testing with commercial software solutions.

– Research and Development: 15’000 CHF

· Identifying the best AI methodology for the case study.

· AI algorithm development.

· Data collection and analysis.

– Prototyping and Evaluation: 5’000 CHF

· Development of initial prototypes.

· Preliminary usability testing and evaluation.

– Project Management and Reporting: 2’000 CHF

· Project coordination and compilation of the feasibility report.

– Generally speaking, how can the Innovation Booster – Fashion & Lifestyle help you?

Fashion industry is very strong in Tessin and such a revolutionary solution would allow the area to be a pioneer in the industry also under a virtual point of view and not only physical.

This may enhance also local schools but for sure will boost the satisfaction of employees, granting the feeling of being part of something new. It has to be said that generally employees of this industry moved around in companies over the area and this will give a chance to power up all of them, possibly attracting even more.

This would be the first step to an application to Innoswiss as the project is ambitious and the 3D modelling for clothing offers many chances of further development and a more democratic one.


Alignment with Sustainable Development Goals (SDGs)

– How does the idea contribute to or align with specific Sustainable Development Goals?

Being able to increment the portion of items that can be easily sewn will support decrease the prototyping needs of the industry, as of today some company have up to five protos travelling all over the globe before approving the item. With a massive 3D production, we will be able to limit this back-and-forth to one final prototype or even eliminate this one.

3D modelling is the future of fashion, still the process today requires long hours’ work that are not compatible with the workload given by physical production.

The longest portion of the process lies in the sewing of the garment that is made by model makers on clothing 3D software, we would like to create an alternative that through AI provides a sewn garments to be perfectioned in existing software.