The main claim is that it is possible to determine intoxication by remote sensing.  This could offer a significant advantage in a certain scope, i.e., in treatment of those with alcoholism, as described in the paper.

There do not appear to be any similar publications in the literature.  The closest reference I can find:

Foss, Robert D., Robert B. Voas, and Douglas J. Beirness. "Using a passive alcohol sensor to detect legally intoxicated drivers." American journal of public health 83.4 (1993): 556-560.

discusses a passive detection method for use at roadblocks that requires the driver to be stopped and is not suitable for surreptitious use.  The paper states that the “instrument draws a sample of exhaled air from in front of the driver's mouth and analyzes it for alcohol content (correcting for the mixture of environmental and exhaled air). Because no mouthpiece or other device touches or enters the body, the use of a passive sensor is not believed to constitute a "search" under the Fourth Amendment and the device can, therefore, be used by police officers for routine screening of drivers.”

Yes.

The results look good.  The supplementary material is very helpful.

No deviations are reported.

The supplementary material should be merged with the main paper in any future version.

There are possible civil rights implications and Fourth Amendment concerns cited in paper above.  This remote sensing method does not require sampling the air in front of a driver, and could potentially be performed without the driver’s knowledge.  However, the purpose as described appears to be to verify performance of other therapies.  A follow-up paper describing such studies would be of interest.

I would like to see more on this project.  Have any larger studies been performed?  What is the maximum effective distance?

None

.

NA

MFC-NIR hyperspectral imaging and speech (imaging more so than speech) can help predict BrAC in a noninvasive way. These claims could alter the way people are tested for sobriety at a single instant making it much easier to stop potentially drunk drivers before they even get into the drivers seat.

Yes, these are novel claims that are demonstrated in the paper.

Yes, many sources of literature where used in the design and application of this experiment.

Yes, the claims are based on the results.

N/A

It seems valid, but there was little, if any, mention of a control. How the data was taken was explicitly stated, but how they knew the BrAC was not clearly stated. 

Be more specific on the sobriety of each subject. It's difficult to determine from the current format. If it were more explicitly stated, it would be easier to understand and grasp. Where they drinking alcoholic beverages during the study? If they were, state it and how the imaging and speech helped capture the decreasing sobriety. If they weren't, maybe have them drink while being studied. It would be a more realistic environment for real people and a better measure of effectiveness.

It's definitely interesting and should be looked into further, but I honestly don't know enough about the field to say it is or isn't "outstanding in its discipline."

While this is a great concept, I'm not sure that it's actually helpful. From what I've gathered, this would be used in popular restuarants/bars to monitor customer's BrAC by an onsite officer in order to prevent drunk driving. That's not very practical in the grand scheme of things. It seems like it would be more efficient for the officers to sit in their patrol car on a popular road taken by patrons of those restuarants/bars. 

I would suggest trying to get this technology to be used at cameras by stoplights. If it could be implemented there, people could watch for potentially drunk drivers and have area patrol cars investigate people that show up drunk on the cameras. 

Good ideas. Interesting topic. Perhaps a different application.

I am a student about to graduate with a chemical engineering degree from an ABET accredited institution.