Depth-guided image processing represents a novel and interdisciplinary frontier in the field of image analysis, where the integration of depth information significantly enhances traditional image processing tasks. Integration of depth information, often captured using 3D sensors, stereo cameras, or LiDAR, is at the heart of the latest advancements in multimodal fusion. 

This workshop seeks to raise the interest in the signal processing community in the potential of multimodal models, leveraging depth information. The workshop ties to the PRIN 2022 “LICAM - AI-powered LiDAR fusion for next-generation smartphone cameras” project, where the depth information captured by Lidar sensors onboard modern smartphones has been successfully exploited to enhance deep image compression and image deblurring models.

The workshop features keynotes and invited talks from distinguished researchers in the image processing field on the topic of multimodal models integrating depth information.

The workshop also features a Challenge on "Lidar-guided Image Deblurring". The challenge will ask participating teams to deblur smartphone images, using Lidar depth maps as guidance and achieving the best visual quality.

Registration

Friday September 12, 13:30-17:50 

[13:30-13:40] Workshop welcome and introduction

---

[13:40-14:10] Invited talk: Prof. Pietro Zanuttigh - "Multimodal understanding"

[14:10-14:35] LICAM project talk: Prof. Diego Valsesia - "Lidar-Guided Image Deblurring"

[14:35-15:00] LICAM project talk: Prof. Alessandro Gnutti - "Lidar-Guided Image Compression"

[15:00-15:30] Invited talk: Prof. Wen-Hsiao Peng - "Enhancing Learned Image Compression for Human and Machine Vision with LiDAR Guidance"

---

[15:50-16:30] The "Lidar-guided Image Deblurring" Challenge

---

Workshop conclusion

Challenge

Click the following link to check all the details on how to participate.