TA4a: Advances in Learning-based Image Processing
Tue, 27 Oct, 08:15 - 09:55 PT (UTC -7)
Location: Room 4
Session Type: Lecture
Session Chair: Guan-Ming Su, Dolby Labs, United States
Track: Speech, Image and Video Processing
Tue, 27 Oct, 08:15 - 08:40 PT (UTC -7)

TA4a.1: SAR-PACT: Tokenization Energy-Aware Complexity Allocation for Practical Learned SAR Image Compression
Zach Button, Paras Maharjan, Abdul Badshah, Zhu Li, University of Missouri-Kansas City, United States
Tue, 27 Oct, 08:40 - 09:05 PT (UTC -7)

TA4a.2: Self-Supervised Seizure Detection in Stereo-EEG
Amirhossein Khalilian-Gourtani, Shira Mack, Elise Van Kwawegen, Peter Rozman, Andrew J. Michalak, Adeen Flinker, New York University, United States
Tue, 27 Oct, 09:05 - 09:30 PT (UTC -7)

TA4a.3: Individualized deepfake detection exploiting the idempotency property of neural networks
Mushfiqur Rahman, Runze Liu, Chau-Wai Wong, Huaiyu Dai, NC State University, United States
Tue, 27 Oct, 09:30 - 09:55 PT (UTC -7)

TA4a.4: Restore-in-Latent: Reference-Guided Diffusion with Text/Face-Aware LoRA for Arbitrary Archival Video Upscaling
Sri Harsha Musunuri, Aupendu Kar, Guan-Ming Su, Dolby Laboratories, Inc., United States