Towards Continual, Online, Unsupervised Depth

02/28/2021
by   Muhammad Umar Karim Khan, et al.
0

Although depth extraction with passive sensors has seen remarkable improvement with deep learning, these approaches may fail to obtain correct depth if they are exposed to environments not observed during training. Online adaptation, where the neural network trains while deployed, with unsupervised learning provides a convenient solution. However, online adaptation causes a neural network to forget the past. Thus, past training is wasted and the network is not able to provide good results if it observes past scenes. This work deals with practical online-adaptation where the input is online and temporally-correlated, and training is completely unsupervised. Regularization and replay-based methods without task boundaries are proposed to avoid catastrophic forgetting while adapting to online data. Experiments are performed on different datasets with both structure-from-motion and stereo. Results of forgetting as well as adaptation are provided, which are superior to recent methods. The proposed approach is more inline with the artificial general intelligence paradigm as the neural network learns the scene where it is deployed without any supervision (target labels and tasks) and without forgetting about the past. Code is available at github.com/umarKarim/cou_stereo and github.com/umarKarim/cou_sfm.

READ FULL TEXT
research
07/17/2022

CULT: Continual Unsupervised Learning with Typicality-Based Environment Detection

We introduce CULT (Continual Unsupervised Representation Learning with T...
research
08/01/2023

Online Prototype Learning for Online Continual Learning

Online continual learning (CL) studies the problem of learning continuou...
research
05/16/2023

Online Continual Learning Without the Storage Constraint

Online continual learning (OCL) research has primarily focused on mitiga...
research
11/25/2022

Unsupervised Continual Semantic Adaptation through Neural Rendering

An increasing amount of applications rely on data-driven models that are...
research
08/21/2023

Color Prompting for Data-Free Continual Unsupervised Domain Adaptive Person Re-Identification

Unsupervised domain adaptive person re-identification (Re-ID) methods al...
research
04/29/2020

Continual Deep Learning by Functional Regularisation of Memorable Past

Continually learning new skills is important for intelligent systems, ye...
research
07/10/2020

Continual Adaptation for Deep Stereo

Depth estimation from stereo images is carried out with unmatched result...

Please sign up or login with your details

Forgot password? Click here to reset