TextBoxes plus plus : A Single-Shot Oriented Scene Text Detector
TextBoxes plus plus : A Single-Shot Oriented Scene Text Detector
复制标题
TextBoxes：面向单次的场景文本检测器
DOI：
10.1109/tip.2018.2825107
复制
发表时间：
2018-08-01
影响因子：
10.6
通讯作者：
Bai, Xiang
中科院分区：
文献类型：
--
作者：
Liao, Minghui;Shi, Baoguang;Bai, Xiang
场景文本检测是场景文本识别系统的重要环节，也是一个具有挑战性的问题。与一般的目标检测不同，场景文本检测的主要挑战在于自然图像中任意方向、小尺寸和显著变化的长宽比的文本。在本文中，我们提出了一个端到端的可训练的快速场景文本检测器，名为TextBoxes++，它可以在单个网络前向传递中以高精度和高效率检测任意方向的场景文本。除了高效的非最大抑制之外，不涉及任何后处理。我们已经在四个公共数据集上评估了拟议的TextBoxes++。在所有实验中，TextBoxes++在文本定位准确性和运行时间方面优于竞争方法。更具体地说，对于1024 x 1024 ICDAR 2015附带文本图像，TextBoxes++在11.6帧/秒时的f-measure为0.817，对于768 x 768 COCO-Text图像，在19.8帧/秒时的f-measure为0.5591。此外，与文本识别器相结合，TextBoxes++在流行的基准测试中显著优于最先进的单词定位和端到端文本识别任务方法。
Scene text detection is an important step of scene text recognition system and also a challenging problem. Different from general object detections, the main challenges of scene text detection lie on arbitrary orientations, small sizes, and significantly variant aspect ratios of text in natural images. In this paper, we present an end-to-end trainable fast scene text detector, named TextBoxes++, which detects arbitraryoriented scene text with both high accuracy and efficiency in a single network forward pass. No post-processing other than efficient non-maximum suppression is involved. We have evaluated the proposed TextBoxes++ on four public data sets. In all experiments, TextBoxes++ outperforms competing methods in terms of text localization accuracy and runtime. More specifically, TextBoxes++ achieves an f-measure of 0.817 at 11.6 frames/s for 1024 x 1024 ICDAR 2015 incidental text images and an f-measure of 0.5591 at 19.8 frames/s for 768 x 768 COCO-Text images. Furthermore, combined with a text recognizer, TextBoxes++ significantly outperforms the state-of-the- art approaches for word spotting and end-to-end text recognition tasks on popular benchmarks.