24小时热门版块排行榜

返回列表

当前主题已经存档。

richards

铁虫 (初入文坛)

应助: 0 (幼儿园)
金币: 35.5
帖子: 3
在线:
虫号: 569949
注册: 2008-06-06
专业: speech recognition

[交流] 请问这个算小修吗?希望大不大？

Dear XXX,

Reviewers have now commented on your paper and, for your guidance, we have appended their comments below. You will see they are advising that you revise your manuscript. If you are prepared to undertake the work required, we would be pleased to reconsider your manuscript for publication.

Reviewers' comments:

Reviewer #1: Key: 1,2 = Page 1, Line 2

This paper is a first revision of an earlier submission. This paper reports on the application of a classification algorithm to normal and pathological voice acoustic data contained within the Kay Elemetrics database of MEEI.

The background of this work are papers published in the IEEE Transactions on Biomedical Engineering (TBME), noted as Reference [12] and [20] in the current submission. These papers and the current authors use identical samples from the MEEI database.

The classification improvement is marginal. Typically, minor improvements to classification are reported in a "Short communication" or "Letters" article within the journal of context, which in this case is IEEE TBME, rather than a full article in a separate journal.

The paper makes many claims that require substantiation or context. Often this can be accomplished by simply providing a reference; in other cases, the claim may need modified or removed. As some examples:

1,38: "Voice diseases have been increasing rapidly due to environmental pollution and unhealthy voice habits."

1,43: "[Laryngoscopy] causes pain and discomfort to the patient."

The entire paper should be carefully examined by the authors in this regard.

2,42: The section is titled: "Speech Data and Feature Extraction" What "feature extraction"?

2,46: sustained vowel /a/ rather than /ah/

2, 49: Should this be 'Kay' databases rather than 'Key'? See also 9,18.

2,51: ".down sampled to 25 kHz for the purpose of this study." Was downsampling performed with or without filtering??

6,22: Should be "In (14),." rather than "In (11)"

Table 3: Omit "Gold Standard"

7,47: Only a pathologist or qualified clinician should diagnose disease. Refer to the results as "machine classification", or some such.

The first two paragraphs of "Discussion and Conclusions" are actually "Results" for the most part.

8,50: The claim "GMM supervectors.are efficient features for the discrimination of normal and pathological voices" should be re-stated. Perhaps the authors mean to say "effective". Efficiency would have to do with the compactness and/or speed of the classification. I suggest "largely effective", since errors remain. See also 9,39.
9,201: "white point"

The References are incomplete. As one example, see [18], where only the year of publication is provided. Ditto for [1], [2], and many others.

Reviewer #2: The revision is much better than the original manuscript. It is easier to read and flows nicely. The abstract needs to be edited for grammar.

麻烦高手帮看看，多谢了！

回复此楼