Box2Poly: Memory-Efficient Polygon Prediction of Arbitrarily Shaped and Rotated Text


METADATA ONLY
Loading...

Date

2024-03-25

Publication Type

Conference Paper

ETH Bibliography

yes

Citations

Altmetric
METADATA ONLY

Data

Rights / License

Abstract

Recently, Transformer-based text detection techniques have sought to predict polygons by encoding the coordinates of individual boundary vertices using distinct query features. However, this approach incurs a significant memory overhead and struggles to effectively capture the intricate relationships between vertices belonging to the same instance. Consequently, irregular text layouts often lead to the prediction of outlined vertices, diminishing the quality of results. To address these challenges, we present an innovative approach rooted in Sparse R-CNN: a cascade decoding pipeline for polygon prediction. Our method ensures precision by iteratively refining polygon predictions, considering both the scale and location of preceding results. Leveraging this stabilized regression pipeline, even employing just a single feature vector to guide polygon instance regression yields promising detection results. Simultaneously, the leverage of instance-level feature proposal substantially enhances memory efficiency (> 50% less vs. the SOTA method DPText-DETR) and reduces inference speed (> 40% less vs. DPText-DETR) with comparable performance on benchmarks. The code is available at https://github.com/Albertchen98/Box2Poly.git.

Publication status

published

Editor

Book title

AAAI-24 Technical Tracks 2

Volume

38 (2)

Pages / Article No.

1219 - 1227

Publisher

AAAI

Event

38th AAAI Conference on Artificial Intelligence (AAAI 2024)

Edition / version

Methods

Software

Geographic location

Date collected

Date created

Subject

CV: Object Detection & Categorization; CV: Scene Analysis & Understanding

Organisational unit

03886 - Schindler, Konrad / Schindler, Konrad check_circle

Notes

Funding

Related publications and datasets