L********e
发帖数: 1 | 1 [D] IBM Zurich Research Plagiarised Our Paper and got it published on CVPR
2022. Is "copy texts" is plagiarism, "copy idea" is not plagiarism?
I am Xianbiao Qi, a computer vision researcher with more than ten years of
research experience. I am writing this blog to complain of a serious case of
deliberate plagiarism of our paper by the employees from聽**IBM Zurich
Research. They did not copy texts, they copied the idea.**
>
>
>Our preprint paper on Arxiv is "Jiaquan Ye, Xianbiao Qi, Yelin He, and etc.
"PingAn-VCGroup's Solution for ICDAR 2021 Competition on Scientific
Literature Parsing Task B: Table Recognition to HTML." arXiv preprint arXiv:
2105.01848, May 2021" and the code was also released.
>
>
>
>Our paper (Ye et al. arXiv: 2105.01848) was plagiarised by a team in IBM
Zurich Research: "**Ahmed Nassar, Nikolaos Livathinos, Maksym Lysak, and
Peter Staar, "TableFormer: Table Structure Understanding with Transformers."
**聽In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern
Recognition, pp. 4614-4623, June 2022."
>
>
>
>Nassar et al. 's paper plagiarized our method, pre-processing, post-
processing, visualization, inference, systematic solution, code, pretrained
models and etc, but did NOT cite our paper in order to not be captured for
the plagiarism.聽 I was notified by many experts in the fields of OCR and
table analysis.
Let me explain the聽plagiarism process step by step.
>**First**,聽聽let me sort out the timeline.
>
>**Our Tablemaster,**
>
>**2021-04-07,**聽we finish the ICDAR 2021 competition-on-scientific-
literature-parsing聽**hosted by another IBM group in Australia**
>
>[https://icdar2021.org/program-2/competitions/competition-on-scientific-
literature-parsing/]https://icdar2021.org/program-2/competitions/
competition-on-scientific-literature-parsing/)
>
>**2021-05-05**, we released our technical report, a preprint聽version, on
Arxiv[https://arxiv.org/pdf/2105.01848.pdf]https://arxiv.org/pdf/2105.01848
.pdf)
>
>**2021-07-29**, we released our source code, and even our slides on Github,
聽[https://github.com/JiaquanYe/TableMASTER-mmocr]https://github.com/
JiaquanYe/TableMASTER-mmocr)
>
>**2021-09**: we released the Tablemaster pretrained model.
>
>**2021-11**聽we even released tablemaster\_mmocr docker environment.
>
>For this project, we were naked. We release all these materials to
facilitate the community, but you plagiarized it.
>
>
>
>**Plagiarized-TableFormer:**
>
>**2022-03-02**, the Plagiarized-TableFormer was released on Arxiv[https://
arxiv.org/pdf/2203.01017v1.pdf]https://arxiv.org/pdf/2203.01017v1.pdf)Their
supplemental聽to cvpr submission.[https://openaccess.thecvf.com/content/
CVPR2022/supplemental/Nassar\_TableFormer\_Table\_Structure\_CVPR\_2022\_
supplemental.pdf]https://openaccess.thecvf.com/content/CVPR2022/
supplemental/Nassar_TableFormer_Table_Structure_CVPR_2022_supplemental.pdf)
>
>**Then,** let me聽highlight聽nine聽clear evidence.聽聽Then I will explain
each piece of evidence in detail.
1. Our methodology, you聽plagiarize
2. Our Pre-processing, you聽plagiarize
3. Our Post-processing, you聽plagiarize
4. Our Inference聽speedup method,聽聽you聽plagiarize
5. Our "tricky" work, you even plagiarize
6. Our text line detection and text line recognition, you聽plagiarize
7. Our systematic solution,聽you聽plagiarize
8. Our visualization,聽聽you聽plagiarize
9. Misleading the audiences in order to not be captured for聽plagiarism |
|
