DeepAI AI Chat
Log In Sign Up

AdaptIS: Adaptive Instance Selection Network

09/17/2019
by   Konstantin Sofiiuk, et al.
SAMSUNG
10

We present Adaptive Instance Selection network architecture for class-agnostic instance segmentation. Given an input image and a point (x, y), it generates a mask for the object located at (x, y). The network adapts to the input point with a help of AdaIN layers, thus producing different masks for different objects on the same image. AdaptIS generates pixel-accurate object masks, therefore it accurately segments objects of complex shape or severely occluded ones. AdaptIS can be easily combined with standard semantic segmentation pipeline to perform panoptic segmentation. To illustrate the idea, we perform experiments on a challenging toy problem with difficult occlusions. Then we extensively evaluate the method on panoptic segmentation benchmarks. We obtain state-of-the-art results on Cityscapes and Mapillary even without pretraining on COCO, and show competitive results on a challenging COCO dataset. The source code of the method and the trained models are available at https://github.com/saic-vul/adaptis.

READ FULL TEXT

page 1

page 2

page 4

page 5

page 6

page 7

page 11

02/14/2020

Layered Embeddings for Amodal Instance Segmentation

The proposed method extends upon the representational output of semantic...
12/02/2017

Recurrent Neural Networks for Semantic Instance Segmentation

We present a recurrent model for semantic instance segmentation that seq...
06/14/2019

Instance Segmentation with Point Supervision

Instance segmentation methods often require costly per-pixel labels. We ...
08/01/2022

Accurate Polygonal Mapping of Buildings in Satellite Imagery

This paper studies the problem of polygonal mapping of buildings by tack...
02/12/2021

Reviving Iterative Training with Mask Guidance for Interactive Segmentation

Recent works on click-based interactive segmentation have demonstrated s...
07/18/2022

Instance-Aware Observer Network for Out-of-Distribution Object Segmentation

Recent work on Observer Network has shown promising results on Out-Of-Di...
07/11/2021

SE-PSNet: Silhouette-based Enhancement Feature for Panoptic Segmentation Network

Recently, there has been a panoptic segmentation task combining semantic...

Code Repositories

adaptis

[ICCV19] AdaptIS: Adaptive Instance Selection Network, https://arxiv.org/abs/1909.07829


view repo