Robust Learning Through Cross-Task Consistency

CVPR 2020 Amir R. Zamir Alexander Sax Nikhil Cheerla Rohan Suri Zhangjie Cao Jitendra Malik Leonidas J. Guibas

Visual perception entails solving a wide set of tasks (e.g., object detection, depth estimation, etc). The predictions made for different tasks out of one image are not independent, and therefore, are expected to be 'consistent'... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK BENCHMARK
Surface Normals Estimation Taskonomy X-TC (Cross-Task Consistency) L1 error 4.80 # 1
Monocular Depth Estimation Taskonomy X-TC (Cross-Task Consistency) L1 error 1.63 # 1
Depth Estimation Taskonomy X-TC (Cross-Task Consistency) L1 error 1.63 # 1

Methods used in the Paper


METHOD TYPE
🤖 No Methods Found Help the community by adding them if they're not listed; e.g. Deep Residual Learning for Image Recognition uses ResNet