Mask-aware IoU for Anchor Assignment in Real-time Instance Segmentation

Date

2021-11-29

Author

ÖKSÜZ, KEMAL
ÇAM, BARIŞ CAN
Kahraman, Fehmi
Baltacı, Zeynep Sonat
Kalkan, Sinan
Akbaş, Emre

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

201
views

0
downloads

This paper presents Mask-aware Intersection-over-Union (maIoU) for assigning anchor boxes as positives and negatives during training of instance segmentation methods. Unlike conventional IoU or its variants, which only considers the proximity of two boxes; maIoU consistently measures the proximity of an anchor box with not only a ground truth box but also its associated ground truth mask. Thus, additionally considering the mask, which, in fact, represents the shape of the object, maIoU enables a more accurate supervision during training. We present the effectiveness of maIoU on a state-of-the-art (SOTA) assigner, ATSS, by replacing IoU operation by our maIoU and training YOLACT, a SOTA real-time instance segmentation method. Using ATSS with maIoU consistently outperforms (i) ATSS with IoU by ~1 mask AP, (ii) baseline YOLACT with fixed IoU threshold assigner by ~2 mask AP over different image sizes and (iii) decreases the inference time by 25% owing to using less anchors. Then, exploiting this efficiency, we devise maYOLACT, a faster and +6 AP more accurate detector than YOLACT. Our best model achieves 37.7 mask AP at 25 fps on COCO test-dev establishing a new state-of-the-art for real-time instance segmentation. Code is available at

URI

https://bmvc2021-virtualconference.com/conference/papers/paper_0031.html
https://hdl.handle.net/11511/96015

Conference Name

The 32nd British Machine Vision Conference

Collections

Department of Computer Engineering, Conference / Seminar

Suggestions

OpenMETU
Core

VHR IMAGE SEGMENTATION USING OVER SEGMENTED REGIONS IN A MRF MODEL Aytekin, Orsan; Rezaeitabar, Yousef; Ulusoy, İlkay (2011-07-29) This paper presents a segmentation method that exploits object based region merging to delineate the context of the image under investigation. There are two main steps. Initially, primitive objects are obtained by morphological operations that generate spectrally homogenous primitives. We assume that primitives are components of semantic objects that are of interest. Next, these primitives therefore are modeled and merged based on expectation maximization. We presented the results of the experiments applied...
Using object-oriented materialized views to answer selection-based complex queries Alhajj, R; Polat, Faruk (1999-09-01) Presented in this paper is a model that utilizes existing materialized views to handle a wide range of complex selection-based queries, including linear recursive queries. Such queries are complex because it is almost impossible for naive users to predict the formulation of their predicate expressions. Object variables bound to objects in the result of a query are allowed to appear in the predicate of that query. Also, the predicate definition is extended to make it possible to have in the output only a sub...
View maintenance in object-oriented databases Alhajj, R; Polat, Faruk (1996-01-01) In this paper, we present a model that facilitates view maintenance within object-oriented databases. For that purpose, we differentiate between two categories of classes, base classes and brother classes. While the former constitute the actual database, the latter are introduced to hold virtual database, i.e., views derived from base classes. To achieve incremental view update, we introduce a modification list into each base class. A series of algorithms are developed to serve the purpose. Finally it happe...
Model-based code generation for HLA federates Adak, Mehmet; Topcu, Okan; Oğuztüzün, Mehmet Halit S. (Wiley, 2010-02-01) This paper addresses the problem of automated code generation for a High Level Architecture compliant federate application given its behavior model. The behavior model is a part of the architectural model of a federation that the federate can participate in. The federate behavior model is based on Live Sequence Charts, adopted as the behavioral specification formalism in the Federation Architecture Metamodel (FAMM). FAMM serves as a formal language for describing federation architectures. An objective is to...
Rank & Sort Loss for Object Detection and Instance Segmentation ÖKSÜZ, KEMAL; ÇAM, BARIŞ CAN; Akbaş, Emre; Kalkan, Sinan (2021-10-17) We propose Rank & Sort (RS) Loss, a ranking-based loss function to train deep object detection and instance segmentation methods (i.e. visual detectors). RS Loss supervises the classifier, a sub-network of these methods, to rank each positive above all negatives as well as to sort positives among themselves with respect to (wrt.) their localisation qualities (e.g. Intersection-over-Union - IoU). To tackle the non-differentiable nature of ranking and sorting, we reformulate the incorporation of error-driven ...

Citation Formats

K. ÖKSÜZ, B. C. ÇAM, F. Kahraman, Z. S. Baltacı, S. Kalkan, and E. Akbaş, “Mask-aware IoU for Anchor Assignment in Real-time Instance Segmentation,” presented at the The 32nd British Machine Vision Conference, İngiltere, 2021, Accessed: 00, 2022. [Online]. Available: https://bmvc2021-virtualconference.com/conference/papers/paper_0031.html.