TiTDet: A tiny text detector with scale-sensitive loss and effective fusion factor

作者
Guangcun Wei,Jihua Fu,Zhifei Pan,Qingge Fang,Zhang Zhi
出处
期刊:Journal of Intelligent and Fuzzy Systems [IOS Press]
卷期号:46 (5-6): 11367-11379
标识
DOI:10.3233/jifs-236317
摘要

The text in natural scenes is often smaller compared to artificially designed text. Due to the small proportion of pixels, low resolution, less semantic information, and susceptibility to complex scenes, tiny text detection often results in many missed detections. To address this issue, this paper draws inspiration from small object detection methods and proposes TiTDet, a detection algorithm more suitable for tiny text. Due to the small proportion of pixels, low resolution, less semantic information, and susceptibility to complex scenes, tiny text detection often results in many missed detections. To address this issue, this paper draws inspiration from small object detection methods and proposes TiTDet, a detection algorithm more suitable for tiny text. Firstly, this paper incorporates a context extraction module and an attention-guided module. These modules guide contextual information learning through a self attention mechanism, while eliminating the possible negative impact caused by redundant information. Regarding multi-scale feature fusion, this paper proposes a fine-grained effective fusion factor, making the fusion process emphasize small object learning more and highlight the feature expression of tiny texts. In terms of post-processing, this paper proposes a differentiable binarization module, incorporating the binarization process into model training. Leveraging the implicit information in the data to drive model improvement can enhance the post-processing effect. Lastly, this paper proposes a scale-sensitive loss, which can handle tiny texts more fairly, fully considering the positional relationship between the predicted and real regions, and better guiding the model training. This paper proves that TiTDet exhibits high sensitivity and accuracy in detecting tiny texts, achieving an 86.0% F1-score on ICDAR2015. The paper also compares the superiority of the method on CTW1500 and Total-Text.

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
1秒前
共享精神的应助被Erpan采纳,获得10
1秒前
1秒前
彭于晏的应助被Lynn采纳,获得10
1秒前
Holic完成签到,获得积分10
2秒前
zyy发布了新的文献求助10
3秒前
4秒前
yy发布了新的文献求助10
4秒前
焦小黑发布了新的文献求助10
5秒前
5秒前
xiaoluoluo完成签到,获得积分10
5秒前
斯文败类的应助被科研小白采纳,获得10
5秒前
终梦的应助被辛酸有很多种采纳,获得20
5秒前
fanfan发布了新的文献求助10
6秒前
chai完成签到,获得积分10
7秒前
zlll完成签到 ,获得积分10
7秒前
乐天完成签到 ,获得积分10
8秒前
大喜完成签到,获得积分10
8秒前
8秒前
水煮电吹风的应助被hsj采纳,获得10
9秒前
xsx1111完成签到,获得积分10
9秒前
cloud云完成签到,获得积分10
10秒前
10秒前
向大闹闹发布了新的文献求助10
11秒前
燕燕完成签到 ,获得积分10
11秒前
12秒前
哇哇大叫发布了新的文献求助10
12秒前
Tang发布了新的文献求助10
13秒前
hsj完成签到,获得积分10
13秒前
13秒前
serendipity完成签到,获得积分10
13秒前
13秒前
流苏发布了新的文献求助10
14秒前
搜集达人的应助被Yutong采纳,获得10
14秒前
15秒前
可可完成签到,获得积分10
15秒前
Erpan发布了新的文献求助10
15秒前
HaoruMeng完成签到,获得积分10
15秒前
打打的应助被zyy采纳,获得30
15秒前
hnu301发布了新的文献求助10
16秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
拉伸高分子熔体中缠结行为的模拟和小角中子散射研究 1700
The World Atlas of Wine 9th edition 1000
奈特基础组织学:配套组织病理学(第4版)(Netter's Essential Histology With Correlated Histopathology) 800
八旗制度の研究 600
Boundaries and Bodhisattvas 500
高分子基功能界面水蒸发材料与应用研究 500
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 工程类 有机化学 化学工程 生物化学 物理 复合材料 内科学 催化作用 光电子学 人工智能 细胞生物学 无机化学 物理化学 数学
热门帖子
关注 科研通微信公众号,转发送积分 8181940
求助须知:如何正确求助?哪些是违规求助? 9571697
关于积分的说明 21819181
捐赠科研通 7778330
什么是DOI,文献DOI怎么找? 3433556
关于科研通互助平台的介绍 2557767
邀请新用户注册赠送积分活动 2479621