What Makes Good Synthetic Training Data for Learning Disparity and Optical Flow Estimation?

01/19/2018
by   Nikolaus Mayer, et al.
0

The finding that very large networks can be trained efficiently and reliably has led to a paradigm shift in computer vision from engineered solutions to learning formulations. As a result, the research challenge shifts from devising algorithms to creating suitable and abundant training data for supervised learning. How to efficiently create such training data? The dominant data acquisition method in visual recognition is based on web data and manual annotation. Yet, for many computer vision problems, such as stereo or optical flow estimation, this approach is not feasible because humans cannot manually enter a pixel-accurate flow field. In this paper, we promote the use of synthetically generated data for the purpose of training deep networks on such tasks.We suggest multiple ways to generate such data and evaluate the influence of dataset properties on the performance and generalization properties of the resulting networks. We also demonstrate the benefit of learning schedules that use different types of data at selected stages of the training process.

READ FULL TEXT

page 4

page 11

page 16

page 17

page 20

page 21

page 22

page 23

research
10/24/2019

Learning Multi-Human Optical Flow

The optical flow of humans is well known to be useful for the analysis o...
research
10/07/2018

Finding Correspondences for Optical Flow and Disparity Estimations using a Sub-pixel Convolution-based Encoder-Decoder Network

Deep convolutional neural networks (DCNN) have recently shown promising ...
research
12/07/2015

A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimation

Recent work has shown that optical flow estimation can be formulated as ...
research
05/08/2021

PCA Event-Based Otical Flow for Visual Odometry

With the advent of neuromorphic vision sensors such as event-based camer...
research
02/20/2018

Uncertainty Estimates for Optical Flow with Multi-Hypotheses Networks

Recent work has shown that optical flow estimation can be formulated as ...
research
05/16/2017

Learning Image Relations with Contrast Association Networks

Inferring the relations between two images is an important class of task...
research
02/13/2018

Robust Deformation Estimation in Wood-Composite Materials using Variational Optical Flow

Wood-composite materials are widely used today as they homogenize humidi...

Please sign up or login with your details

Forgot password? Click here to reset