Computer Science > Computer Vision and Pattern Recognition
[Submitted on 23 Apr 2020 (v1), last revised 26 May 2020 (this version, v2)]
Title:Location-Aware Feature Selection Text Detection Network
View PDFAbstract:Regression-based text detection methods have already achieved promising performances with simple network structure and high efficiency. However, they are behind in accuracy comparing with recent segmentation-based text detectors. In this work, we discover that one important reason to this case is that regression-based methods usually utilize a fixed feature selection way, i.e. selecting features in a single location or in neighbor regions, to predict components of the bounding box, such as the distances to the boundaries or the rotation angle. The features selected through this way sometimes are not the best choices for predicting every component of a text bounding box and thus degrade the accuracy performance. To address this issue, we propose a novel Location-Aware feature Selection text detection Network (LASNet). LASNet selects suitable features from different locations to separately predict the five components of a bounding box and gets the final bounding box through the combination of these components. Specifically, instead of using the classification score map to select one feature for predicting the whole bounding box as most of the existing methods did, the proposed LASNet first learn five new confidence score maps to indicate the prediction accuracy of the bounding box components, respectively. Then, a Location-Aware Feature Selection mechanism (LAFS) is designed to weightily fuse the top-$K$ prediction results for each component according to their confidence score, and to combine the all five fused components into a final bounding box. As a result, LASNet predicts the more accurate bounding boxes by using a learnable feature selection way. The experimental results demonstrate that our LASNet achieves state-of-the-art performance with single-model and single-scale testing, outperforming all existing regression-based detectors.
Submission history
From: Zilin Wang [view email][v1] Thu, 23 Apr 2020 06:45:26 UTC (9,371 KB)
[v2] Tue, 26 May 2020 02:26:21 UTC (9,619 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.