Irony alert: If you’re going to write a paper about face recognition technology, well, do we really need to go on?
A group of researchers in Wuhan, China, evidently didn’t quite realize they were walking into a ridicule trap when they agreed to have their paper, “Face Recognition with Learning-based Descriptor,” published in IERI Procedia. The article appeared in 2012 and was part of an issue devote to that year’s International Conference on Future Computer Supported Education, which took place in Seoul.
And now comes this:
This article has been retracted at the request of the Author.
The authors have plagiarized part of a paper that had already appeared in [CVPR, IEEE (2010) 2707–2714. http://dx.doi.org/101109/CVPR.2010.5539992]. One of the conditions of submission of a paper for publication is that authors declare explicitly that their work is original and has not appeared in a publication elsewhere. Re-use of any data should be appropriately cited. As such this article represents a severe abuse of the scientific publishing system. The scientific community takes a very strong view on this matter and apologies are offered to readers of the journal that this was not detected during the submission process.
We’re curious about the “part of a paper” bit here. We found the plagiarized article, and judging by the title, at least, it seems like the two papers are awfully close. Indeed, the titles are identical, which ought to have been a red flag for the editors of IERI Procedia. Here’s the abstract of the lifted article:
We present a novel approach to address the representation issue and the matching issue in face recognition (verification). Firstly, our approach encodes the micro-structures of the face by a new learning-based encoding method. Unlike many previous manually designed encoding methods (e.g., LBP or SIFT), we use unsupervised learning techniques to learn an encoder from the training examples, which can automatically achieve very good tradeoff between discriminative power and invariance. Then we apply PCA to get a compact face descriptor. We find that a simple normalization mechanism after PCA can further improve the discriminative ability of the descriptor. The resulting face representation, learning-based (LE) descriptor, is compact, highly discriminative, and easy-to-extract. To handle the large pose variation in real-life scenarios, we propose a pose-adaptive matching method that uses pose-specific classifiers to deal with different pose combinations (e.g., frontal v.s. frontal, frontal v.s. left) of the matching face pair. Our approach is comparable with the state-of-the-art methods on the Labeled Face in Wild (LFW) benchmark (we achieved 84.45% recognition rate), while maintaining excellent compactness, simplicity, and generalization aability across different datasets.
We think there must be a lesson in there somewhere.
Utterly trivial, pedantic whine follows. “Procedia”? This is the second recent post involving a journal with a fake, pseudo-Latin name. “Procedia” isn’t even close. After 50 years, I can’t remember what the ***ing proper construction is — “processa” or “procedenda” or some such — but it sure isn’t “procedia.” Have some pity on the poor language. Latin is dead already. Publishers don’t need to abuse it any more. Nihil nisi bonum, and all that.
Actually “Procedia” is not a verb but a noun, precisely the plural nominative of “Procedium” meaning “Proceeds” (i.e. Yield)… I presume that “proceeds” derives from the same verb (“to proceed”) of the gerunde “proceedings”.
The “literal” translation of Proceedings in latin would be “Procedens”… but the “correct” translation is “Acta”
Harrrrumph. But there is no such noun in the Latin corpus. That’s the whole point. One would have to use the verb procedere (3rd conjugation?) as the passive participle (processus: that which has been advanced) or perhaps a gerund-ish form (procedendus?: that which should/must be advanced — but I’m guessing on that one). Both of these are verb constructions used to create nouns or adjectives. Giving the publisher the benefit of the doubt on using the neuter, that gives you processum (singular), or processa (plural). If you’d had Mr. Wischert as your Sixth Form Latin master, you’d know better.
Title language issues aside, it is common in the field of computer vision to publish a ‘journal version’ of work presented at a conference (which is what CVPR is). However, conference submissions are not abstracts; in the case of CVPR it is a ten-page (or twelve, don’t quite remember) full paper. Additionally, to get into some such conferences (including CVPR) the quality of the paper has to be so high that sometimes it precludes later publication in a journal because of excessive overlap. It’s a pity.
Why is that a pity? The conference proceedings are, at least to me, an equal part of the scientific literature. They are fully citeable, searchable, whatsoever… It is only a pity for people that long for duplicate publications.