Timezone: »
This demonstration showcases our novel innovations on self-supervised monocular depth estimation. First, we enhance self-supervised monocular depth estimation with semantic information during training. This reduces the error by 12% and achieves state-of-the-art performance. Second, we enhance the backbone architecture using a scalable method for neural architecture search which optimizes directly for inference latency on a target device. This enables operation at > 30 FPS. We demonstrate these techniques on a smartphone powered by a Snapdragon® Mobile Platform.
Author Information
Hong Cai (Qualcomm AI Research)
Yinhao Zhu (Qualcomm AI Research)
3D vision, generative modeling, neural data compression
Janarbek Matai (Qualcomm AI Research)
Fatih Porikli (Qualcomm)
Fatih Porikli is an IEEE Fellow and a Senior Director at Qualcomm. He was a full tenured Professor in the Research School of Engineering, Australian National University (ANU), Canberra. He served as the Vice President of San Diego CBG Device Hardware Competency Center, Futurewei, San Diego. He was the Chief Scientist of Autonomous Vehicles at Futurewei, Santa Clara. Until 2017, he was the Computer Vision Research Group Leader at Data61/CSIRO, Australia (before merging, at NICTA). He has received his Ph.D. from New York University (NYU), New York, in 2002. Previously he served a Distinguished Research Scientist at Mitsubishi Electric Research Laboratories (MERL), Cambridge. Before joining MERL in 2000, he developed satellite imaging solutions at HRL, Malibu CA, and 3D display systems at AT&T Research Labs, Middletown, NJ. He has contributed broadly to object and motion detection, object tracking, image enhancement and super-resolution, visual representations, and video analytics. His current research interests include computer vision, pattern recognition, deep learning, manifold learning, sparse optimization, multimedia processing, data analytics, and online learning with many commercial applications, including efficient AI platforms, mobile phones, AI/VR, video surveillance, medical systems, automotive perception, car navigation, intelligent transportation, logistics, satellite systems, automation, visualization, and consumer electronics. Fatih Porikli was the recipient of the R&D 100 Scientist of the Year Award in 2006. He has won 6 best paper awards at premier conferences, including the IAPR MVA Test-of-Time Best Paper of the Decade in 2019, the Best Paper on Deep/Machine Learning at APSIPA 2017, the Best Student Paper at IEEE ACCV 2016, the Best Poster Award at IEEE AVSS in 2014, the Best Paper at IEEE AVSS in 2011, the Best Paper at IEEE Workshop on Object Tracking and Classification Beyond Visible Spectrum in 2010, and the Best Paper Runner-Up at IEEE CVPR in 2007. He has received 7 other professional prizes. He authored more than 300 publications, co-edited two books, and invented 100+ US patents. He served on the organizing committees of several flagship conferences, including CVPR, ICCV, ECCV, ICRA, IROS, AAAI, ICIP, AVSS, ICPR, AI, ICME, ISVC, and ICASSP. He was the General Chair and Technical Program Chair of the IEEE Winter Conference on Applications of Computer Vision (WACV) in 2014 and 2015, and IEEE Advanced Video and Signal based Surveillance Conference (AVSS) in 2010 and 2012. He has been an Associate Editor of several IEEE and Springer journals for the past 15 years. He was a panelist judge at several NSF proposal panels from 2006 to 2013 and gave keynotes in various events. He supervised more than 60 Ph.D. students. His h-index: 66, i10-index: 226, citation count: 21,000+
Fei Yin (Qualcomm AI Research)
Tushar Singhal (Qualcomm)
Bharath Ramaswamy (Qualcomm Technologies Inc.)
Frank Mayer (Qualcomm Inc.)
Chirag Patel (Qualcomm)
Parham Noorzad (Qualcomm Technologies, Inc.)
Andrii Skliar (Qualcomm AI Research)
Tijmen Blankevoort (Qualcomm)
Joseph Soriaga (Qualcomm Technologies, Inc.)
Ron Tindall (Qualcomm)
Pat Lawlor (Qualcomm)
More from the Same Authors
-
2021 : Spatial-Temporal Gated Transformersfor Efficient Video Processing »
Yawei Li · Babak Ehteshami Bejnordi · Bert Moons · Tijmen Blankevoort · Amirhossein Habibian · Radu Timofte · Luc V Gool -
2023 Poster: OpenShape: Scaling Up 3D Shape Representation Towards Open-World Understanding »
Minghua Liu · Ruoxi Shi · Kaiming Kuang · Yinhao Zhu · Xuanlin Li · Shizhong Han · Hong Cai · Fatih Porikli · Hao Su -
2022 Expo Demonstration: Full-Stack 3D Scene Understanding on an Extended Reality Headset »
Hong Cai -
2022 Expo Demonstration: Conditional Compute for On-device Video Understanding »
Tijmen Blankevoort -
2021 : Unsupervised Indoor Wi-Fi Positioning »
Farhad G. Zanjani · Ilia Karmanov · Hanno Ackermann · Daniel Dijkman · Max Welling · Ishaque Kadampot · Simone Merlin · Steve Shellhammer · Rui Liang · Brian Buesker · Harshit Joshi · Vamsi Vegunta · Raamkumar Balamurthi · Bibhu Mohanty · Joseph Soriaga · Ron Tindall · Pat Lawlor -
2021 Social: Shine in Your Technical Presentation »
Armina Stepan · Tijmen Blankevoort -
2020 Poster: Bayesian Bits: Unifying Quantization and Pruning »
Mart van Baalen · Christos Louizos · Markus Nagel · Rana Ali Amjad · Ying Wang · Tijmen Blankevoort · Max Welling