CVPR 2018 Kicks Off; Best Papers Announced

The 2018 Conference on Computer Vision and Pattern Recognition (CVPR) opened today in Salt Lake City, USA. The CVPR organizing committee used the occasion to announce its coveted Best Paper and Best Student Paper selections.

One of the world’s top three academic conferences in the field of computer vision (along with ICCV and ECCV), CVPR this year received 3,300 main conference paper submissions and accepted 979. Over 6,500 will attend the five-day conference, which features 70 oral presentations, 224 spotlight orals, posters, 21 tutorials, 48 workshops, an annual doctoral consortium, and the conference’s biggest-ever industrial exhibition with over 115 participating companies.

Taking top honours in the Best Paper category is Taskonomy: Disentangling Task Transfer Learning from a Stanford University and University of California at Berkeley research group. The paper proposes a computational approach for finding correlations between different tasks by using transfer learning techniques to solve new tasks with reduced dependency on labeled data.

The paper’s first author is Postdoctoral Researcher at Stanford University and UC Berkeley Amir Roshan Zamir, who told Synced “this paper basically brings the concept of transfer learning to machine learning and perception on how you can solve one problem that you face with what you have already known.”

The Best Student Paper award meanwhile goes to Total Capture: A 3D Deformation Model for Tracking Faces, Hands, and Bodies from a Carnegie Mellon University research group. This paper presents a unified deformation model for the markerless capture of multiple scales of human movement, including facial expressions, body motion, and hand gestures.

Also announced this morning were CVPR Honorable Mentions for the papers: Deep Learning of Graph Matching, SPLATNet: Sparse Lattice Networks for Point Cloud Processing, CodeSLAM — Learning a Compact, Optimisable Representation for Dense Visual SLAM, and Efficient Optimization for Rank-based Loss Functions.

Meanwhile, to recognize young researchers’ distinguished contributions to computer vision, the IEEE Pattern Analysis and Machine Intelligence (PAMI) Technical Committee presented the PAMI Young Researcher Award to Andreas Geiger, a University of Tübingen professor, and Kaiming He, a research scientist at Facebook AI Research (FAIR).

Andreas Geiger (centre), Kaiming He (right)

The PAMI Technical Committee also awarded the PAMI Longuet-Higgins Prize to A Discriminatively Trained, Multiscale, Deformable Part Model, a paper from CVPR 2008, in recognition of its enduring relevancy over a 10-year period.

Synced has identified a number of other facts and trends at CVPR this year that readers might be interested in:

  • There were 1715 reviewers assigned to evaluate submitted papers for CVPR 2018, more than the total number of CVPR attendees just four years ago.
  • The CVPR organizing committee received over US$2 million in sponsorships.
  • Google AI Research Scientist Jordi Pont-Tuset suggested in his blog that Generative Adversarial Networks (GANs) might catch up with deep learning someday. More than eight percent of CVPR 2018’s accepted papers include “GANs” in their titles, doubling the frequency at CVPR 2017.
  • Meanwhile, Long short-term memory (LSTMs) models are finding it difficult to keep up.

CVPR 2018 runs to June 22, and Synced will be reporting from the conference throughout the week.

Journalist: Tony Peng | Editor: Michael Sarazen

Follow us on Twitter @Synced_Global for more AI updates!

Subscribe here to get insightful tech news, reviews and analysis!

Let’s talk about AI and FinTech! Synced invites you to join our DTalk Episode One: Deploying AI in Mobile-First Customer-facing Financial Products: A Tale of Two Cycles. Jike Chong will share his ideas on employing AI techniques in FinTech business model. Register at! We hope to see you on June 21st in Silicon Valley.