Overview

This tutorial covers topics at the frontier of research on visual recognition and its use in higher-level tasks (the beyond). The first three talks cover methods and principles behind image classification, object detection, instance segmentation, and semantic segmentation. After the coffee break, the final two talks explore newly emerging tasks involving the use of visual recognition for action and reasoning.

Schedule

Room 355 ABC, Friday June 22, 2018 9am - 12:00pm

09:00 - 09:30. Learning Deep Representations for Visual Recognition - Kaiming He [slides]

09:30 - 10:00. The Generalized R-CNN Framework for Object Detection - Ross Girshick [slides]

10:00 - 10:30. Panoptic Segmentation: Unifying Semantic and Instance Segmentations - Piotr Dollár [slides]

10:30 - 11:00. Coffee Break

11:00 - 11:30. Embodied Vision - Georgia Gkioxari [slides]

11:30 - 12:00. Scene Graphs for Recognition and Generation - Justin Johnson [slides]