3 Takeaways from ICCV 2019

The workshops and presentations during the International Conference on Computer Vision 2019 (ICCV) were a great preview into what’s next in computer vision.Here are a few highlights from the premier international computer vision event for students, academics and industry researchers.

PHOTO: Sama Team at ICCV 2019

Key Takeaways from ICCV 2019

• ICCV attendance increased by 150 percent from 2017 to 2019, going from 3,000 attendees to 7,500. Likewise, ICCV received more than 4,000 papers, up by 50 percent since 2017. Everything from convolutional neural networks to neural architecture search, to robust object detection was published. This article highlights the best paper awards received.
• On the exhibit floor, we found varied applications of computer vision, from a pair of cameras detecting a person's face, age, and emotional state to algorithms helping doctors detect breast cancer more easily. There were also some interesting poster presentations, some of which were highlighted in the ICCV Daily 2019 such as, “Align, Attend and Locate: Chest X-ray Diagnosis via Contrast Induced Attention Network with Limited Supervision.”

PHOTO: Exhibitor booth at ICCV 2019

• Throughout the conference, opinions ran high around the use of lidar and radar vs camera sensors only for autonomous driving. There were various approaches on display outlining specific use cases for different sensor packages.

PHOTO: Qualcomm exhibitor display at ICCV 2019

ICCV 2019 provided a welcoming platform for the distribution and discussion of scholarly and technical work, and the Sama team looks forward to continuing conversations on how to securely train and validate computer vision.