• Home
  • Search
  • Toward Joint Thing-and-Stuff Mining for Weakly Supervised Panoptic Segmentation
  • Cite Icon15
  • https://doi.org/10.1109/cvpr46437.2021.01642Copy DOI Icon

Toward Joint Thing-and-Stuff Mining for Weakly Supervised Panoptic Segmentation

  • Jun 1, 2021
  • Yunhang Shen +8 more
Show More
  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Panoptic segmentation aims to partition an image to object instances and semantic content for thing and stuff categories, respectively. To date, learning weakly supervised panoptic segmentation (WSPS) with only image-level labels remains unexplored. In this paper, we propose an efficient jointly thing-and-stuff mining (JTSM) framework for WSPS. To this end, we design a novel mask of interest pooling (MoIPool) to extract fixed-size pixel-accurate feature maps of arbitrary-shape segmentations. MoIPool enables a panoptic mining branch to leverage multiple instance learning (MIL) to recognize things and stuff segmentation in a unified manner. We further refine segmentation masks with parallel instance and semantic segmentation branches via self-training, which collaborates the mined masks from panoptic mining with bottom-up object evidence as pseudo-ground-truth labels to improve spatial coherence and contour localization. Experimental results demonstrate the effectiveness of JTSM on PASCAL VOC and MS COCO. As a by-product, we achieve competitive results for weakly supervised object detection and instance segmentation. This work is a first step towards tackling challenge panoptic segmentation task with only image-level labels.

Similar Papers
  • Conference Article
  • Citations4

CASNet: Common Attribute Support Network for image instance and panoptic segmentation

  • Jan 10, 2021
  • Xiaolong Liu +4
  • Conference Article
  • Citations7

Learning Panoptic Segmentation from Instance Contours

  • May 30, 2021
  • Sumanth Chennupati +4
  • Conference Article
  • Citations2

Dual-CNN fusion panoptic segmentation based on edge optimization

  • Feb 15, 2022
  • Yaling Li +2
  • Conference Article

Common Methods of Image Panoptic Segmentation Based on Deep Learning

  • Jan 22, 2021
  • Congcong Wang
  • Research Article
  • Citations172

Leveraging Instance-, Image- and Dataset-Level Information for Weakly Supervised Instance Segmentation.

  • Sep 10, 2020
  • IEEE Transactions on Pattern Analysis and Machine Intelligence
  • Yun Liu +5
  • Research Article

Panoptic Segmentation for Indoor Environments using MaskDINO: An Experiment on the Impact of Contrast

  • Jan 28, 2025
  • ELCVIA Electronic Letters on Computer Vision and Image Analysis
  • Khalisha Putri +1
  • Conference Article
  • Citations1

Contextual Image Parsing via Panoptic Segment Sorting

  • Oct 20, 2021
  • Jyh-Jing Hwang +2
  • Research Article
  • Citations220

Few-Example Object Detection with Model Communication.

  • Jun 07, 2018
  • IEEE Transactions on Pattern Analysis and Machine Intelligence
  • Xuanyi Dong +4
  • Research Article
  • Citations100

Weakly Supervised Large Scale Object Localization with Multiple Instance Learning and Bag Splitting.

  • Feb 01, 2016
  • IEEE Transactions on Pattern Analysis and Machine Intelligence
  • Weiqiang Ren +3
  • Conference Article
  • Citations2

LDA-AQU: Adaptive Query-guided Upsampling via Local Deformable Attention

  • Oct 28, 2024
  • Zewen Du +4
  • Conference Article
  • Citations348

Weakly Supervised Complementary Parts Models for Fine-Grained Image Classification From the Bottom Up

  • Jun 01, 2019
  • Weifeng Ge +2
  • Conference Article
  • Citations49

Part-aware Panoptic Segmentation

  • Jun 01, 2021
  • Daan De Geus +4
  • Research Article
  • Citations5

Attention-effective multiple instance learning on weakly stem cell colony segmentation

  • Jan 24, 2023
  • Intelligent Systems with Applications
  • Novanto Yudistira +3
  • Conference Article

Revisiting milis multiple instance learning algorithm with a different instance selection mechanism

  • Apr 01, 2012
  • Osman Akin +2
  • Conference Article

FIMIL : A high-throughput deep learning model for abnormality detection with weak annotation in microscopy images

  • Feb 04, 2020
  • Proceedings of the Australasian Computer Science Week Multiconference
  • Jing Ke +5
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.