Continual Semantic Segmentation via Repulsion-Attraction of Sparse and Disentangled Latent Representations

03/10/2021
by   Umberto Michieli, et al.
10

Deep neural networks suffer from the major limitation of catastrophic forgetting old tasks when learning new ones. In this paper we focus on class incremental continual learning in semantic segmentation, where new categories are made available over time while previous training data is not retained. The proposed continual learning scheme shapes the latent space to reduce forgetting whilst improving the recognition of novel classes. Our framework is driven by three novel components which we also combine on top of existing techniques effortlessly. First, prototypes matching enforces latent space consistency on old classes, constraining the encoder to produce similar latent representation for previously seen classes in the subsequent steps. Second, features sparsification allows to make room in the latent space to accommodate novel classes. Finally, contrastive learning is employed to cluster features according to their semantics while tearing apart those of different classes. Extensive evaluation on the Pascal VOC2012 and ADE20K datasets demonstrates the effectiveness of our approach, significantly outperforming state-of-the-art methods.

READ FULL TEXT

page 4

page 7

page 15

page 16

page 17

page 18

page 19

page 20

research
11/23/2020

PLOP: Learning without Forgetting for Continual Semantic Segmentation

Deep learning approaches are nowadays ubiquitously used to tackle comput...
research
06/23/2021

Multiband VAE: Latent Space Partitioning for Knowledge Consolidation in Continual Learning

We propose a new method for unsupervised continual knowledge consolidati...
research
04/11/2023

Continual Semantic Segmentation with Automatic Memory Sample Selection

Continual Semantic Segmentation (CSS) extends static semantic segmentati...
research
09/29/2022

RECALL: Rehearsal-free Continual Learning for Object Classification

Convolutional neural networks show remarkable results in classification ...
research
05/25/2023

Fairness Continual Learning Approach to Semantic Scene Understanding in Open-World Environments

Continual semantic segmentation aims to learn new classes while maintain...
research
06/01/2023

Continual Learning for Abdominal Multi-Organ and Tumor Segmentation

The ability to dynamically extend a model to new data and classes is cri...
research
01/09/2023

CaSpeR: Latent Spectral Regularization for Continual Learning

While biological intelligence grows organically as new knowledge is gath...

Please sign up or login with your details

Forgot password? Click here to reset