Hello From Perú, This apps is available in some library?. Is there some workpaper about it? Is there some web site to read? I'm interested in this subject. Thank you for the response and congratulation
@yoshiocalderon Thanks for your interest! We are getting ready to implement our first test of the real system in the Art and Architecture library at Miami University. We will be doing an experiment to determine how much it improves speed and accuracy of shelf reading, and hope to present results at either LITA or ACRL. When the time comes, we will post some updated videos.
I am very interested in how this could be used to keep video materials in order. If you got it to the point where the QR code on a DVD style case could be read would the other QR codes need to be of the same size? I am wondering because the collection I manage has many DVDs, but there still is a huge volume of VHS tapes and I'm wondering how would the app deal with multiple size QR codes. This is very exciting work!
@thevideolibrarian It should not be too hard to allow both DVD-sized cases and VHS-sized cases on the same shelf. Do you put the call number on the spine of the DVD, right now? With our current tag-printing functionality, the call number wouldn't fit on something that narrow, but the tag itself would fit...
I've often felt it would be nice to have an app that recognizes book spines based in image artwork and text and the like - so that you more could easily upload your book collection to a shared book service such as librarything - rather than having to present the front of each book to a recognizer ... and rather than having to put a special marker image onto each book.
This is fabulous! Now, if you could only add a cheap way to get a GPS reading on the books... probably a privacy violation, but my, it would certainly be helpful for those of us who work with little library patrons who regularly misplace their books!
I'm with skudfisher, once this is working, adding a robot to do both the shelf reading and moving the books into their correct positions would be great.
This is neat. How many books must be in order for the app to work? For example, if every other book were out of order (so, one correct, one wrong, one correct, one wrong, etc), would the app still work?
@alexlent Good question. Our algorithm identifies the minimum number of moves necessary to sort the books, and then it only marks the books that need to be moved with a red x. So yes, it will "work" no matter what. On the other hand, if half of the books were wrong, then it would probably be easier/faster to just sort them by hand.
This is neat. How many books must be in order for the system to work? For example, if every other book were out of order, would the system still work?
All this needs now is a little autonomous robot that wanders the stacks at night creating a photo report of books that have been wrongly shelved for librarians to arrange in the morning :) I worked in a library for 5 years and shelf reading is one of the most tedious and boring things I've ever done in my life. This is awesome!
This is super cool. Will this work with library collections that are still growing? How are new books integrated into the shelf read function? Libraries continuously add books to their collections.
@debhussein The code on the spine directly encodes the call number of the book, so you don't have to do anything to the application to keep it up to date, and you don't need internet access on your smartphone (after the app has been installed). All you have to do is put your call number into our web page, and it will generate a spine tag for you. Stick it on the book, and you are done.
@ShuriRyuKa Sorry, the application is still under development, and is not available for other people to try yet. If everything goes well we will be soliciting for beta-test partners some time around Christmas.
I have to wonder how you would deal with books who's spine is not wide enough to hold an AR tag? Also how will the AR tags scale up to deal with larger collections? You can't really repeat the use of the tags across the full collection or you would run the risk of the app flagging something that is horribly out of place as ok.
@cap60552 Very good questions. This is a prototype which only uses 16 bits on the tag. We are working on a version that allows us to put ~72 bits per tag. This should mean it will scale up to work with any collection.
Items narrower than 1/4" will be problematic. We use the "video" mode of the camera, which only gives us 28 pixels per inch when scanning a whole shelf. You can make the tags smaller if you are willing to get closer, and scan only half a shelf or 1/3 of a shelf at a time.
@DrBoOhio the 1/3 of the shelf seems like it would be usable for the average student worker. Perhaps even a panning scan of each shelf would be useful, with an auditory beep or some other tone if an out of place item is seen.
I also find myself daydreaming about this technology for a slightly different application. While scanning the shelf, one could easily gather GPS coordinates of each item. Combine that with Google Maps in your OPAC and provide realtime item locating for patrons.
I noticed each book has an distinct AR tag so I'm assuming that each tag corresponds to a title. So if this system was deployed in a library, what is the means to add newly acquired books to a shelf?
@liquidsorcery JediKnightMike is correct, the tag corresponds to the LC call number. We intend to eventually support the other call number systems, but right now it is LC only. We will eventually have a simple web-based tool where you just paste in a list of LC numbers, one per line, and it generates the tags for you as a PDF or PNG.
The application is not actually as slow as it appears in the video. The slow frame-rate is due to the way I am capturing the video ... by taking screen-shots over and over.
Hello From Perú, This apps is available in some library?. Is there some workpaper about it? Is there some web site to read? I'm interested in this subject. Thank you for the response and congratulation
yoshiocalderon 1 week ago
@yoshiocalderon Thanks for your interest! We are getting ready to implement our first test of the real system in the Art and Architecture library at Miami University. We will be doing an experiment to determine how much it improves speed and accuracy of shelf reading, and hope to present results at either LITA or ACRL. When the time comes, we will post some updated videos.
DrBoOhio 1 week ago
@DrBoOhio Ok thanks!!! I joined your canal youtube. Good luck!!!. I will be follow you.
yoshiocalderon 1 week ago
I am very interested in how this could be used to keep video materials in order. If you got it to the point where the QR code on a DVD style case could be read would the other QR codes need to be of the same size? I am wondering because the collection I manage has many DVDs, but there still is a huge volume of VHS tapes and I'm wondering how would the app deal with multiple size QR codes. This is very exciting work!
thevideolibrarian 4 months ago
@thevideolibrarian It should not be too hard to allow both DVD-sized cases and VHS-sized cases on the same shelf. Do you put the call number on the spine of the DVD, right now? With our current tag-printing functionality, the call number wouldn't fit on something that narrow, but the tag itself would fit...
DrBoOhio 4 months ago
All seems well. however what happens when a misplaced book is not on the same shelf and in the screen?
kristis61 7 months ago
press 0 not too fast
MrCutterTV 10 months ago
press 0 not too ast
MrCutterTV 10 months ago
ZOMG I LOVE YOU For making this!! I am sick and tired of trying to find what I need in the library and finding everything out of place!
trivialinformation 10 months ago
this is for lazy people who can't use there own brain to work out the book should go...
insideaNWO 10 months ago
Dude, you should be proud.
Good job!
Illimitus 10 months ago
my wife, a librarian, just had a book-gasm...
oakXXIII 10 months ago 4
Excuse me, but do you really need a gadget to do this?
Serai3 10 months ago
I've often felt it would be nice to have an app that recognizes book spines based in image artwork and text and the like - so that you more could easily upload your book collection to a shared book service such as librarything - rather than having to present the front of each book to a recognizer ... and rather than having to put a special marker image onto each book.
anselmhook 10 months ago
This is fabulous! Now, if you could only add a cheap way to get a GPS reading on the books... probably a privacy violation, but my, it would certainly be helpful for those of us who work with little library patrons who regularly misplace their books!
msnokleby 10 months ago
You are right, this is my hot librarian fantasy.
iamthebestartist 10 months ago
@iamthebestartist
hehe mine too
trivialinformation 10 months ago
I'm with skudfisher, once this is working, adding a robot to do both the shelf reading and moving the books into their correct positions would be great.
connieblogger 11 months ago
@connieblogger Ah, mind you if a robot is doing the shelving, you likely won't need to do shelf-reading. ;)
connieblogger 11 months ago
Sorry about the double post.
alexlent 11 months ago
This is neat. How many books must be in order for the app to work? For example, if every other book were out of order (so, one correct, one wrong, one correct, one wrong, etc), would the app still work?
alexlent 11 months ago
@alexlent Good question. Our algorithm identifies the minimum number of moves necessary to sort the books, and then it only marks the books that need to be moved with a red x. So yes, it will "work" no matter what. On the other hand, if half of the books were wrong, then it would probably be easier/faster to just sort them by hand.
DrBoOhio 11 months ago
This is neat. How many books must be in order for the system to work? For example, if every other book were out of order, would the system still work?
alexlent 11 months ago
All this needs now is a little autonomous robot that wanders the stacks at night creating a photo report of books that have been wrongly shelved for librarians to arrange in the morning :) I worked in a library for 5 years and shelf reading is one of the most tedious and boring things I've ever done in my life. This is awesome!
skudfisher 11 months ago
This is super cool. Will this work with library collections that are still growing? How are new books integrated into the shelf read function? Libraries continuously add books to their collections.
debhussein 11 months ago
@debhussein The code on the spine directly encodes the call number of the book, so you don't have to do anything to the application to keep it up to date, and you don't need internet access on your smartphone (after the app has been installed). All you have to do is put your call number into our web page, and it will generate a spine tag for you. Stick it on the book, and you are done.
DrBoOhio 11 months ago
@DrBoOhio I checked your web page and didn't see a place to input the LC call number. Could you post the URL here? Thanks!
ShuriRyuKa 10 months ago
@ShuriRyuKa Sorry, the application is still under development, and is not available for other people to try yet. If everything goes well we will be soliciting for beta-test partners some time around Christmas.
DrBoOhio 10 months ago
I am very impressed!
I have to wonder how you would deal with books who's spine is not wide enough to hold an AR tag? Also how will the AR tags scale up to deal with larger collections? You can't really repeat the use of the tags across the full collection or you would run the risk of the app flagging something that is horribly out of place as ok.
cap60552 11 months ago
@cap60552 Very good questions. This is a prototype which only uses 16 bits on the tag. We are working on a version that allows us to put ~72 bits per tag. This should mean it will scale up to work with any collection.
Items narrower than 1/4" will be problematic. We use the "video" mode of the camera, which only gives us 28 pixels per inch when scanning a whole shelf. You can make the tags smaller if you are willing to get closer, and scan only half a shelf or 1/3 of a shelf at a time.
DrBoOhio 11 months ago
@DrBoOhio the 1/3 of the shelf seems like it would be usable for the average student worker. Perhaps even a panning scan of each shelf would be useful, with an auditory beep or some other tone if an out of place item is seen.
I also find myself daydreaming about this technology for a slightly different application. While scanning the shelf, one could easily gather GPS coordinates of each item. Combine that with Google Maps in your OPAC and provide realtime item locating for patrons.
cap60552 11 months ago
Besides the obvious (capture method problem) this is fucking awesome. Mad props, and I'd love to see more.
sudoscientist 11 months ago
I noticed each book has an distinct AR tag so I'm assuming that each tag corresponds to a title. So if this system was deployed in a library, what is the means to add newly acquired books to a shelf?
liquidsorcery 11 months ago
@liquidsorcery I assume the AR tag corresponds to the LoC call number.
JediKnightMike 11 months ago
@liquidsorcery JediKnightMike is correct, the tag corresponds to the LC call number. We intend to eventually support the other call number systems, but right now it is LC only. We will eventually have a simple web-based tool where you just paste in a list of LC numbers, one per line, and it generates the tags for you as a PDF or PNG.
DrBoOhio 11 months ago
The application is not actually as slow as it appears in the video. The slow frame-rate is due to the way I am capturing the video ... by taking screen-shots over and over.
DrBoOhio 11 months ago 23
@DrBoOhio it really kills the demonstration, though. awesome work otherwise!
Deject3d 11 months ago