Shelving Made Easy (or Easier)

April 6, 2011, 4:18 pm

Putting misshelved books back in their proper places is not a library worker’s favorite task. It takes time and it’s not exactly scintillating. Now a computer-science professor has come up with a way to make the process faster and less burdensome: an augmented-reality shelf-reading app that can scan an entire shelf’s worth of books at a time and alert workers which ones are out of place.

Bo Brinkman is an associate professor of computer science and software engineering at Miami University, in Ohio. A specialist in augmented reality and computer ethics, he happens to be married to the university’s art-and-architecture librarian. Hearing his wife talk about trying to motivate student workers to do more shelf-reading got Mr. Brinkman thinking about creative solutions to the problem.

The app he came up with, tentatively called Shelvar, relies on special tags—kind of like QR codes—attached to the books’ spines. Each tag “exactly represents the call number” of each book, Mr. Brinkman explains. A user with a current-generation smartphone or tablet computer scans the shelf using the app, and Shelvar indicates which books aren’t in the right places. Visual cues, including directional arrows, indicate where the misfiled book ought to go.

There are inventory apps that scan books one at a time, according to Mr. Brinkman, but he doesn’t know of any others that work with many volumes at once. “Our hope is to be a hundred times faster—instead of scanning one book at a time, you scan 144 books at a time,” he says.

The biggest hassle with Shelvar involves tagging books in the first place. For new acquisitions, libraries could just add the tags as they catalog and process each book. The next step is figuring out just how big a deal it is to tag existing collections. In the next round of testing, “the main thing we’re going to be looking at is how much time does it take to put on the tags, and how much time does it save once they’re on?” Mr. Brinkman says.

He designed most of the code. Matt Hodges, his undergraduate research assistant, focused on the user interface. For the prototype, they worked with Google Nexus One smartphones and Samsung Galaxy tablet. For the moment, the app works on the Android platform, but Mr. Brinkman intends to design an iPhone version as well. “There’s no way we can market it successfully if we don’t have both,” he says.

Just how it’s marketed—or if it’s marketed at all—depends on several factors. First, Mr. Brinkman has to find out how well the app really works and whether it’s the time- and-money-saver he hopes it will be. He plans to do an alpha test late this year in one or two sections of Miami University’s library system, involving maybe 70 to 100 linear feet of books and one student worker. If that goes well, he’ll stage a much bigger trial.

Then the university and the state “have first rights to patent anything,” the computer-science professor says. “If the school decides to pass on it, we’ll probably have a very different business model.”

At the moment, Mr. Brinkman thinks that the shelf-reading app probably ought to be free, so that library workers could just download it on their phones, requiring a minimal investment upfront. More sophisticated versions might find a market, either among libraries or even commercial entities such as booksellers. “In the end, it only makes sense if it’s a lot cheaper than the way they’re currently doing it,” he says. “Otherwise they wouldn’t buy it anyway.”

Even at this early stage, Shelvar has turned some heads. It made a splash at the recent Association of College and Research Libraries meeting in Philadelphia. “I would say that the biggest reaction I’ve gotten has been, ‘Wow, it would be nice not to have to do shelf-reading,’” Mr. Brinkman says. The YouTube demo video he posted on March 22 has gotten more than 30,000 hits. Augmented-reality enthusiasts have shown even more interest than librarians have, he says, because so much augmented-reality design hasn’t really focused on practical needs yet.

For librarians, though, Shelvar could be the answer to a very practical concern. “Keeping the books in order is not what librarianship is all about, but it’s something they have to do in order to give users a good experience,” Mr. Brinkman says.

This entry was posted in Libraries, Mobile College Apps. Bookmark the permalink.

E-mail

jffoster

Must have been some mistake. The TCU is the Horned Frogs, not the Horny Frogs.
willynilly

If it is true that TCU has some mystical mental power and can actually identify and reject “miscreant” athletes at the admissions level, then we are left with only one remaining explanation. Something at TCU, something horrible in fact, transforms these outstanding, moral, high purpose male/female athletes and totally corrupts their moral compasses, turning them into stalking rapists. TCU has a real problem on its hands.
jffoster

On their hands indeed.

Warts.
old nassau’67

I’d like to know (a) what sport and (b) what “criminal records”?
On one hand (in view of comments 2 and 3 above, dare I use this metaphor?): “TCU Football Recognized For Its Graduation Rate…./TCU one of just four programs to be ranked in the top 25 in the AP poll and APR”
But, on the other, “The university’s graduation rates for its athletes in all but one sport lag behind the national averages, but university officials said they hope new academic facilities and programs for athletes will help them improve.
Chris Uchacz, director of athletic academic services, said TCU’s overall graduation success rate or GSR, is 68 percent. The NCAA’s national average is 78 percent.”
richman

As I said on Twitter, the sheer possibilities of this app are amazing, anything that you put on a shelf could be checked/counted in almost no time. Physical inventory/searching may never be the same.
etmiller

When I worked in the main library stacks at UC Berkeley, this task, known as “fine ordering,” was the most disagreeable that one could be assigned. Moreover, it was so cumbersome, and the collection was so large–and perused so much–that one was always plagued by the existential awareness that the stacks would perpetually be in need of it–fine ordering would never be completely accomplished.. Now, perhaps, that has changed–and I will one day truly rest at ease!
etmiller

I would add, that I’m curious how the “man-versus-machine” challenge pans out. Until the hand-held processor and screen refresh rate are faster, I wonder if it’s even quicker than a human–(a human that’s really paying attention). And can it spot labels that are partially obscured or incorrectly?
extra88

Timing of this seems wrong. If you’re going to the trouble of handling every book, it should be for something that technology isn’t going to outgrow quickly. Machine vision is improving rapidly enough that a phone will be able to simply OCR the spines to read whatever title and author information is there as well as existing cataloging stickers.

While more costly, I think RFIDs in books would have a longer “shelf life.” Phones (or dedicated hardware) will be NFC devices that you can “wand” over a shelf to read the RFID cards then augmented reality could be used as Shelvar does to highlight which books are out of order. The RFIDs could also be used at Circulation (though I suppose these QR codes could be as well) but they could also be used at door security in lieu of the magnetic strips used today.
http://twitter.com/bgrigsby24 Bryon Grigsby

Incredible app. Imagine the time saved that could be more productive.
http://www.facebook.com/profile.php?id=100000194160038 Judy Quist

I don’t understand how this app deals with books that are too skinny to have a label/tag on the spine? In this clip all the books are thicker, and I’m wondering how you would get the visual on the thinner materials in your collection?
dilmore

I saw the demonstration at ACRL last week and was interested in the idea, particularly as I worked as a student assistant for three years and spent many hours shelf reading. However I left the session wondering if the camera on a smart phone could capture the image of an entire shelf of books in the close quarters presented by typical library book stacks. I tried it after I returned to my library and found that I could not capture an image of an entire shelf even in the areas of the library where shelving units are farther apart than they are in most of our collection. I have software on my smart phone that provides limited telephoto capacity, but I have not found anything that allows wide angle image capture. It will be interesting to see how this works in the alpha test.
http://twitter.com/chriskox Christopher Kox

It really is mind numbing work and no human should be assignrd to more than an hour shift. That said, it has its pleasures, and it provides employment for students, or high school graduates. This does increase the number of identifying attributes for each item and in the form of yet another label turning the book into a flagship for paste-ons. Finally, libraries using RFIDs may already have adequate technology for rapid reading, and external bar codes facilitate fast scanning for inventory. Still, if this technology can streamline inventory, it may have value, but our institution has found inventory ever more demanding since automated. On this, cheap labor, performed manually was less trying.
harringcbh

How about integrating this app with “Delicious Library”? I’d love to see both together!
from, forever trying to locate a book in my libraries,
clemg

Libraries that have RFID systems have been able to do this for years.
joansey

I’d be interested to see a similar application that incorporates OCR of call numbers, eliminating the need to relabel an entire collection. I know OCR can be glitchy, but it always seems to be improving… Besides, who isn’t a little glitchy after hours of shelf reading? :)
bfrank1

25 years too late. Shelves are emptying out as materials are moved to storage. This would have been useful in the ’80′s, when libraries first began coding and relabeling for automated circulation. No one would go to all the effort to do this again today, when we aren’t sure we will even have books much longer. If you wanted to do it, the only remotely effective way would be to just start, and wait until the number of tagged books was sufficient to make it work – plus, the question about books with labels on the side is a very real problem – that’s where a lot of the errors and hard work issues are anyway.
not4nothin

Guess maybe his wife didn’t know about RFID.

The downside of either system would be the need to re-tag all the items in the collection. Unless you’re doing a major project like changing classification schemes from Dewy to LC, and you’ll be handling every item in the collection anyway, remarking the entire collection would be very expensive – prohibitively so in most cases.

In the end, the library keeps limping along doing the same things it has done in the past. So what if they lose some stuff because it’s miss-shelved? It’s far cheaper to simply make the poor student workers suffer with the mind numbing work of shelf reading.

File it under manageable mismanagement.
rivenhomewood

Putting labels in books is nothing new for libraries – we all did it to all our books just a few years ago. It was called “retrospective conversion” or “recon” and involved pasting a barcode label in each book so we could automate our checkout process. I worked at one library where we did the whole collection twice, a little at a time, and at another where we got it all done in a week. If this ap is successful, academic libraries have a lot of cheap student labor available for the initial set-up.

In recon, the main expense was not putting tags in books but getting your cataloging data into machine-readable form and printing a tag that was keyed to each book. Using this app would probably mean scanning a tag, using a computer connection (which could be a smart phone) to associate it with the book’s catalog record, then applying the tag to the book – fairly time-consuming but quite possible. It’s possible that the expense of doing it might be less than the ongoing expense of training and employing student shelfreaders.

However, as one commenter points out, it might be more cost-effective to just put an RFD tag in each book, and then use that both for check out and shelf reading. That would also take care of the problem posed by books too thin to have a tag on their spine.
mbelvadi

What’s the detection distance of rfid? I would have thought that it couldn’t be used for this because several inches of books would all “read” simultaneously and not give any useful information about their relative order.
http://www.facebook.com/people/Emily-Byron/1484965443 Emily Byron

I think we officially need this at our library!
harmsjb

As a former shelver, I had the same question: How would it deal with skinny books? If this (or any) technology’s means of “data input” is only visual, then any volumes without enough outward-facing surface area might present a problem. Perhaps the program could simply ignore those volumes and focus (no pun intended) only on volumes that do have an easily readable tag/label. Anything else might have to be checked the old-fashioned way, i.e. by a human being.

As others have commented, whether it’s worth the time and effort to (re)label every book in the library would need to be considered. And I’m skeptical that current smartphone/tablet cameras are (yet) up to the task of doing effective OCR on the motley assortment and quality of spine labels of various ages found in a typical academic library. However, Shelvar is still in its early stages, so perhaps it could address some of these issues.

So while it seems unlikely to be a complete solution, something like Shelvar could, under the right conditions, make a good portion of the shelfreading process go a lot faster overall.
609zr

The following is a quote from an article entitled Education: The Ph.D. Factory. “Scientists who attain a PhD are rightly proud — they have gained entry to an academic elite. But it is not as elite as it once was. The number of science doctorates earned each year grew by nearly 40% between 1998 and 2008.” In the U.S. and Japan, supply exceeds demand. China and India graduate an abundance of Ph.D.s, but “the quality of the graduates is not consistent.” From a supply and demand perspective, the Indian government would benefit from incentives to bring its Ph.D.s home.
makers34

Re-title this article, “It’s Time to End Universities”, and I’d say you’re onto something.
chramirez

I work with a variety of students as I teach college preparatory math, and teach at my community college part time. I also substitute teach and tutor math students in high school. I understand my students’ frustrations with the different expectations of college due to being in their highschool classrooms and seeing firsthand the different ways math is taught as compared to a college classroom. In many highschool classrooms, each student is issued a graphing calculator, and they solve problems involving polynomials and linear equations with the calculator. There is much emphasis on analyzing the product of their calculations and less on the steps to reaching a solution. The benefit might be more high level thinking and less being stuck on the steps. But when they come to us they may in some cases lack some skills and have less patience for learning these steps on paper since we have no graphing calculators and only allow simple function calculators in the classroom.