Skip to Main content Skip to Navigation
Master thesis

YOLO-Based Panoptic Segmentation

Manuel Alejandro Diaz-Zapata 1
1 CHROMA - Robots coopératifs et adaptés à la présence humaine en environnements dynamiques
Inria Grenoble - Rhône-Alpes, CITI - CITI Centre of Innovation in Telecommunications and Integration of services
Abstract : Given the recent challenge of Panoptic Segmentation, where every pixel in an image must be given a label, as in semantic segmentation, and an instance id, a new YOLO-based architecture is proposed here for this computer vision task. This network uses the YOLOv3 architecture, plus parallel semantic and instance segmentation heads to perform full scene parsing. A set of solutions for each of these two segmentation tasks are proposed and evaluated, where a Pyramid Pooling Module is found to be the best semantic feature extractor given a set of feature maps from the Darknet-53 backbone network. The network gives good segmentation results for both stuff and thing classes by training with a frozen backbone, where boundaries between background classes are consistent with the ground truth and the instance masks match closely the true shapes of the objects present in a scene.
Complete list of metadata

https://hal.inria.fr/hal-02884735
Contributor : Manuel Diaz Zapata Connect in order to contact the contributor
Submitted on : Tuesday, February 9, 2021 - 3:19:52 PM
Last modification on : Wednesday, November 3, 2021 - 3:58:53 AM

File

MSc_Thesis_HAL_v2.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02884735, version 2

Collections

Citation

Manuel Alejandro Diaz-Zapata. YOLO-Based Panoptic Segmentation. Computer Vision and Pattern Recognition [cs.CV]. 2020. ⟨hal-02884735v2⟩

Share

Metrics

Les métriques sont temporairement indisponibles