Astri Logo White
Search icon
Astri Logo White
search icon

Deep Learning Based Real Time Computer-Generated Holograms (CGH) for SLM projection (ARD/334)

Project Title:
Deep Learning Based Real Time Computer-Generated Holograms (CGH) for SLM projection (ARD/334)
Project Reference:
ARD/334
Project Type:
Seed
Project Period:
27 / 12 / 2024 - 26 / 12 / 2025
Funds Approved (HK$’000):
2,780.700
Project Coordinator:
Mr Derek LIU
Deputy Project Coordinator:
/
Deliverable:
Research Group:
/
Sponsor:
Description:

Augmented reality projection applications such as digital lighting, heads-up displays (HUD), and screen-less displays are becoming increasingly important in the trend towards next generation smart vehicles; they will play an important role in visual communication for all road users. The traditional spatial light modulation (SLM) optical system of computer-generated holograms (CGH) has problems such as color resolution mismatch and the slow operation of the CGH iterative algorithm, which poses challenges in market application. Due to the introduction of SSC (Spatially Separated Color) independent reconstruction of light paths and hologram calculation technology based on deep learning, the customized laser projection system we proposed can effectively increase the light path output and reduce CGH generation time to adapt to real-time applications. In this project, the R&D team will develop an RGB discrete optical path compensation structure optical engine based on the deep learning CGH algorithm for laser-based real-time projection display. This technology is superior to existing traditional amplitude projection systems in terms of light efficiency and 3D image generation capabilities of augmented reality images, which will bring a differentiated and practical solution to automotive HUD and smart projection applications.

Co-Applicant:
/
Keywords:
/