Product design sketch understanding: unsupervised learning, neural rendering, and novel 3D shape representations
The goal of this PhD proposal is to exploit and advance AI and deep learning to facilitate the creative design process.
Start date1 July 2022
Funding sourceSurrey Institute for People-Centred AI
A stipend of £15,609 for 2021/22, which will increase each year in line with the UK Research and Innovation (UKRI) rate, plus Home rate fee allowance of £4,500 (with automatic increase to UKRI rate each year). For exceptional international candidates, there is the possibility of obtaining a scholarship to cover overseas fees
Ideation and concept sketching is used by product designers to externalize a shape from their imagination. At the ideation stage designers create multiple quick sketches. At the concept sketching stage, to achieve an accurate shape depiction, designers use dedicated sketching techniques in which earlier strokes create anchors for future strokes and carefully depict shape proportions in perspective. Concept sketching is the central element of the product design pipeline, usually followed by prototyping 3D shapes as digital or physical models. To facilitate a quick understanding of the 3D shape, the PhD work will address the problem of lifting 2D sketches to 3D in an unsupervised fashion.
The goal of this PhD proposal is to exploit and advance AI and deep learning to facilitate the creative design process. The primary objective is to exploit multi-view sketching and to propose new 3D shape representations that are easy for designers to work with. Leveraging state-of-art in differentiable and neural rendering the successful applicant will develop unsupervised methods aiming at 3D reconstruction. The student will have the opportunity to collaborate with industrial designers to collect an extended dataset of concept and/or ideation sketches to demonstrate the effectiveness of the proposed approach on representative product design sketches. The synthetic data and existing sketches from the OpenSketch dataset will be used for prototyping deep models.
In summary, the PhD work is expected to:
- Introduce multi-view shape encoders that leverage perspective distortions and different levels of details
- Introduce new 3D shape representations in the context of deep learning, and network architectures enabling such representations
- Contribute a new dataset of multi-view concept sketches.
Related linksSurrey Institute for People-Centred Artificial Intelligence Pixelor: a competitive sketching AI agent (PDF) Graphical convention formation during visual communication (PDF)
All applicants should have (or expect to obtain) a first-class degree in a numerate discipline (mathematics, science or engineering) or MSc with Distinction (or 70% average) and a strong interest in pursuing research in this field. Passion for sketching and product design is advantageous.
English language requirements
IELTS Academic 6.5 or above (or equivalent) with 6.0 in each individual category.
How to apply
Applications should be submitted via the Centre for Vision, Speech and Signal Processing programme page. Please clearly state the studentship title and supervisor on your application.