文豆 & 文库

隆峰2112:

经典的构图和调色。

天朗气清:

- You are my girl.

- I will always be your girl.

岸边望海:

晓莲:

生命当如泉水般透明,清澈,那便是人的本性的复苏。


点击欣赏:生活

 

點滴的記錄:

深圳华为坂田基地占地1.3平方公里,其次,由培训中心、研发中心、中试中心、行政中心、生产中心、数据中心、机加中心及百草园员工公寓等各具特色的建筑群组成,为员工提供了一个优美、整洁、舒适的工作环境。具体来说,包括行政中心A区,华电B区,数据中心C区,研发中心E区和F区,生产中心G区,技加中心K区,培训中心J区。


Tumblr:
https://essayjeans.tumblr.com/post/185425762764/huawei

點滴的記錄:

湛天雲海碧波影:

ICDAR 2019 Robust Reading Challenge on Scanned Receipts OCR and Information Extraction Results


SROIE Introduction:

Scanned receipts OCR is a process of recognizing text from scanned structured and semi-structured receipts, and invoices in general. On the other hand, extracting key texts from receipts and invoices and save the texts to structured documents can serve many applications and services, such as efficient archiving, fast indexing and document analytics. Scanned receipts OCR and information extraction (SROIE) play critical roles in streamlining document-intensive processes and office automation in many financial, accounting and taxation areas. However, SROIE also faces big challenges. With performance greatly boosted by recent breakthroughs in deep learning technologies in terms of accuracy and processing speed, OCR is becoming mature for many practical tasks (such as name card recognition, license plate recognition and hand-written text recognition). However, receipts OCR has much higher accuracy requirements than the general OCR tasks for many commercial applications. And SROIE becomes more challenging when the scanned receipts have low quality. Therefore, in the existing SROIE systems, human resources are still heavily used in SROIE. There is an urgent need to research and develop fast, efficient and robust SROIE systems to reduce and even eliminate manual work.


 

Task 1 - Scanned Receipt Text Localisation
Hmean Ranking Table
Rank Team Name Team Members Insititute Recall Precision Hmean
1 SCUT-DLVC-Lab-Refinement Jiapeng Wang*, Yan Li*, Tianwei Wang, Jiaxin Zhang, Yichao Huang, Canjie Luo, Kai Ding, Lianwen Jin (*equal contribution) South China University of Technology, INTSIG Information Co. Ltd 98.64% 98.53% 98.59%
2 Ping An Property & Casualty Insurance Company Xianbiao Qi, Ning Lu, Yuan Gao, Yihao Chen, Shaoqiong Chen, Wenwen Yu, Rong Xiao Ping An Property & Casualty Insurance Company 98.60% 98.40% 98.50%
3 H&H Lab HUST_VLRGROUP(Mengde Xu, Zhen Zhu, Hui Zhang, Mingkun Yang, Jiehua Yang) & HUAWEI_CLOUD_EI(Jing Wang, Yibin Ye, Shenggao Zhu, Dandan Tu) Huazhong University of Science and Technology & Huawei Technologies Co. Ltd Joint Laboratory 97.93% 97.95% 97.94%
4 Psenet_dcn Xiufeng Jiang - 96.62% 96.21% 96.42%
5 BOE_IOT_AIBD v5 BOE_IOT_AIBD - 95.95% 95.99% 95.97%
6 EM_ocr Hao Wu, Na He, Zhou Shen, Dan Meng, Qingfeng Wang - 95.85% 96.08% 95.97%
7 Clova OCR Seung Shin, Sungrae Park, Seonghyeon Kim, Jaeheung Surh, Junyeop Lee, Hwalsuk Lee Clova AI Research, NAVER Corp 96.04% 95.79% 95.92%
8 IFLYTEK-textDet_v3 IFLYTEK IFLYTEK 93.77% 95.89% 94.81%
9 A Single-Shot Model for Robust Text Localization Hanqin Wang, Jie Qin, Fan Zhu, Li Liu, and Ling Shao Inception Institute of Artificial Intelligence 93.93% 94.80% 94.37%
10 SituTech_OCR Kui Lyu, Tianhao Tang, Minghao Wang SituTech 93.81% 94.18% 94.00%

 


Task 2 - Scanned Receipt OCR
Hmean Ranking Table
Rank Team Name Team Members Insititute Recall Precision Hmean
1 H&H Lab HUST_VLRGROUP(Hui Zhang, Mingkun Yang, Mengde Xu, Zhen Zhu, Jiehua Yang) & HUAWEI_CLOUD_EI(Jing Wang, Yibin Ye, Shenggao Zhu, Dandan Tu) Huazhong University of Science and Technology & Huawei Technologies Co. Ltd Joint Laboratory 96.35% 96.52% 96.43%
2 HeReceipt-Ensemble Yichao Huang*, Tianwei Wang*, Jiaxin Zhang*, Yan Li, Jiapeng Wang, Canjie Luo, Kai Ding, Lianwen Jin (*equal contribution) INTSIG Information Co. Ltd, South China University of Technology 94.56% 95.10% 94.82%
3 Ping An Property & Casualty Insurance Company Xianbiao Qi, Yihao Chen, Shaoqiong Chen, Ning Lu, Yuan Gao, Wenwen Yu, Rong Xiao Ping An Property & Casualty Insurance Company 94.48% 94.86% 94.67%
4 CLOVA OCR Sungrae Park, Seung Shin, Seonghyeon Kim, Jaeheung Surh, Junyeop Lee, Hwalsuk Lee Clova AI Research, NAVER Corp 94.30% 94.88% 94.59%
5 SCUT-DLVC-Lab-Lexicon Tianwei Wang*, Jiaxin Zhang*, Yichao Huang*, Jiapeng Wang, Yan Li, Canjie Luo, Kai Ding, Lianwen Jin (*equal contribution) South China University of Technology, INTSIG Information Co. Ltd 94.18% 94.88% 94.53%
6 DenseNet-Attention Recognition PINGAN Tech PINGAN Tech 94.29% 94.58% 94.44%
7 CITlab Argus Text Recognition Tobias Grüning, Gundram Leifert, Jochen Zöllner, Tobias Strauß, Roger Labahn CITlab 93.55% 93.61% 93.58%
8 Unet followed by CRNN with CTC Roberto Lotufo, Ramon Pires - 88.58% 87.30% 87.93%
9 BOE_IOT_AIBD T2 V5 BOE_IOT_AIBD - 87.84% 86.66% 87.24%
10 CRNN after UNet Segmentation Roberto Lotufo, Ramon Pires, Israel Campiotti, Rubens Machado, Luis Serrano, Giovanni Garuffi - 85.77% 86.48% 86.12

 


Task 3 - Key Information Extraction from Scanned Receipts
Hmean Ranking Table
Rank Team Name Team Members Insititute Recall Precision Hmean
1 Ping An Property & Casualty Insurance Company Xianbiao Qi, Wenwen Yu, Ning Lu, Yihao Chen, Shaoqiong Chen, Yuan Gao, Rong Xiao Ping An Property & Casualty Insurance Company 90.49% 90.49% 90.49%
2 EAST det + Multi-class classification liyulin, v_huangju, xiequnyi, qinxiameng Baidu 89.70% 89.70% 89.70%
3 H&H Lab HUST_VLRGROUP(Hui Zhang, Mengde Xu, Mingkun Yang, Zhen Zhu, Jiehua Yang) & HUAWEI_CLOUD_EI(Jing Wang, Yibin Ye, Shenggao Zhu, Dandan Tu) Huazhong University of Science and Technology & Huawei Technologies Co. Ltd Joint Laboratory 89.63% 89.63% 89.63%
4 CLOVA OCR Sungrae Park, Seonghyeon Kim, Seung Shin, Jaeheung Surh, Junyeop Lee, Hwalsuk Lee Clova AI Research, NAVER Corp 89.05% 89.05% 89.05%
5 NiuBiHongHong Ge daye - 87.61% 87.61% 87.61%
6 HeReceipt-withoutRM Hanmin Duan, Zhiqin Lu, Yang Chang, Yan Li, Yichao Huang, Kai Ding INTSIG Information Co. Ltd, South China University of Technology 83.00% 83.24% 83.12%
7 BOE_IOT_AIBD_v3 BOE_IOT_AIBD - 82.71% 82.71% 82.71%
8 PATECH_CHENGDU_OCR JunKun Zhou, Ming Guan, ZhengNan Luo, MingTao Wang, YuBin Xiao, MingBin Hou - 81.70% 82.29% 82.00%
9 NER with spaCy model Roberto Lotufo, Giovani Garuffi, Israel Campiotti, Ramon Pires, Rubens Machado, Luis Serrano - 78.96% 79.02% 78.99%
10 CITlab Argus Information Extraction (positional & line features, enhanced gt) Tobias Strauß, Tobias Grüning, Gundram Leifert, Jochen Zöllner, Roger Labahn CITlab 77.38% 77.38% 77.38%

 

Result Link:
http://www.onlyou.com/sroie/index.html

 

云南境内最大的高原湖泊滇池,素有“高原明珠”之称,也是鸟类的天堂。1988年,滇池以昆明滇池风景名胜区的名义,被国务院批准列入第二批国家级风景名胜区名单。每年冬天,来自西伯利亚的候鸟红嘴海鸥都会随着季节的变化来到滇池过冬,它们大约每年11月飞来昆明,来年3月离开返回西伯利亚。湖面上数以万记的红嘴鸥成为滇池冬季的一大景观,成了春城昆明的一张美丽名片。 那里游人如织,成千上万的海鸥漫天飞舞、大有遮天蔽日之势,非常壮观。 隆冬时节的春城昆明,阳光和煦、清风拂面。著名观鸟地滇池海埂大坝人山人海,热闹非凡。

 

汤博乐:
https://essaybeans.tumblr.com/post/185425982343/dianchi


新加坡南洋理工大学小笼包。不仅仅外观极具代表性和冲击力,其内部构造也很吸睛。这座建筑共8层,提供了56个没有拐角、没有明确前后之分的教室,据说可供33000名学生同时使用(@essaybeans

无住生心:

虎啸山林:

揽月听风:

1X摄影精选:

原图地址:https://gallery.1x.com/photo/1672451/

澳大利亚摄影师Richard Vandewalle的作品《Kyoto Night》。夜晚,古老的街道上空无一人,白天拥挤的人群早已散去。摄影师独自行走在京都祇园角,仿佛回到了过去。整张作品安静而想和,适当的柔焦更加增添了这种气氛。画面左边树叶上的杂光可能稍微有些多余。


七彩夫人:

qiangr1209:

susuxihuanlehu:

【人生6条智慧】1、知而不言,谨记言多必失;2、自我解脱,才能自我超越;3、冷静思考,才能让自己更清醒;4、用心看世界,才会看清人的本来面目;5、放下,放下了才能重新开始;6、感恩,才能在逆境中寻求希望。

 (图文/网络 编辑/susuxihuanlehu)