Retraction Note: IoT-based 3D convolution for video salient object detection

Shizhou Dong, Zhifan Gao, Sandeep Pirbhulal, Gui‐Bin Bian, Heye Zhang, Wanqing Wu, Shuo LiNeural 2024 · DOI · OpenAlex

Abstract

This report examines the formal retraction of the 2019 article "IoT-based 3D convolution for video salient object detection," originally published in the journal Neural Computing and Applications. Following a comprehensive investigation by Springer Nature, the article was retracted due to systemic failures in the editorial and peer-review processes. This analysis details the nature of the compromised editorial handling and the subsequent impact on the scientific record.

Introduction

In July 2024, the academic community was alerted to the formal retraction of a study proposing an IoT-based 3D convolution method for video salient object detection. While the original 2019 paper aimed to contribute to the field of computer vision, its validity has been entirely rescinded. The retraction serves as a critical correction, highlighting the ongoing challenges journals face in maintaining editorial integrity, particularly within specialized publication formats.

The Investigation: Compromised Editorial Integrity

The original article was part of a guest-edited special issue, a format intended to spotlight emerging or niche research areas. However, the publisher's investigation revealed that this specific issue was subject to "compromised editorial handling." This term indicates that the standard "gatekeeping" functions—designed to ensure that only rigorous, relevant, and peer-reviewed research is published—were either bypassed or manipulated.

Key findings from the investigation included:

  • Peer Review Failures: The integrity of the peer-review process was found to be fundamentally compromised. This means the technical claims of the paper never received the objective scrutiny required for scientific publication.
  • Citation Irregularities: The presence of inappropriate or irrelevant references suggested the paper's involvement in "citation rings." These are coordinated efforts to artificially inflate the impact metrics of specific authors or journals through unnecessary citations.
  • Scope Misalignment: The content was identified as potentially being out of scope for the journal or the specific theme of the guest-edited issue, further suggesting a breakdown in editorial oversight.

Consequences for the Scientific Record

The primary "result" of this retraction notice is the total nullification of the original study. The Editor-in-Chief of Neural Computing and Applications has explicitly stated a loss of confidence in the results and conclusions of the work. Consequently, the technical claims regarding 3D convolution and IoT frameworks in the original paper (DOI: 10.1007/s00521-018-03971-3) are no longer considered reliable.

The "Check for updates" (CrossMark) icon is a standard tool in academic publishing that allows researchers to verify if a paper has been retracted or updated.

Furthermore, the publisher, Springer Nature, has adopted a "neutral" stance regarding jurisdictional claims and institutional affiliations mentioned in the original work. This effectively distances the journal and the publisher from the paper's history and any potential institutional disputes.

Challenges in Post-Publication Correction

The retraction process also highlighted the logistical difficulties of policing the scientific record years after a paper's initial publication. The publisher was unable to obtain current contact information for four of the seven authors (Dong, Gao, Pirbhulal, and Wu). Among the authors who were reached, there was a notable lack of consensus: author Heye Zhang officially disagreed with the retraction, while others did not respond. This underscores the complex legal and professional disputes that often accompany the removal of published research.

Conclusion

This retraction is part of a broader, ongoing effort by major publishers to utilize forensic tools and internal investigations to identify and remove "paper mill" products and compromised research from the global body of knowledge. For researchers in the fields of IoT and computer vision, the primary takeaway is the necessity of verifying the status of older citations. The "Check for updates" system remains a vital tool for ensuring that future research is built upon a foundation of verified and high-integrity data.

Summary updated

BibTeX
@inproceedings{dong2024retraction,
  author = {Shizhou Dong and Zhifan Gao and Sandeep Pirbhulal and Gui‐Bin Bian and Heye Zhang and Wanqing Wu and Shuo Li},
  title = {{Retraction Note: IoT-based 3D convolution for video salient object detection}},
  booktitle = {Neural},
  year = {2024},
  doi = {10.1007/s00521-024-10072-x},
}

Loading related content...

Figure