Those of you close to me, know that I have a camera and photo fetish. In fact, I’m now on like my 10th digital camera. This coupled with my “scan everything I have” project from a few years ago, has amassed into a image library of over 55,000 files.
Early in my digital photo days, I tried every conceivable way to organize the images. I tried folders for people, naming the photos, folders for events, and various software programs. Every method had one or more of following faults.
- It was based on EXIF information (Exchangeable Image File Format)
- Doesn’t account for Multiple Objects/Situations/People on single images
- Requires separate index of images
1 – EXIF Information
In my opinion the best photography website on the Internet is Flickr. It make extremely good use use of EXIF information. Flickr is great for your “best” photos you want to share, but it does not lend itself to a repository for images in the numbers I’m working with.
A few thousand images in my collection are scanned from prints, which contain no EXIF information, except for the manufacturer and model of the scanner. The image files do not reflect the date of the photograph without manually tweaking this information. Doing so would take a ton of time, but undoubtably it is a task I will have to tackle one day.
2 – Multiple Objects/Situations/People
Quite a few photography blogs suggest renaming the images to reflect the date, subject, location, etc that relate to it. Imaging renaming that family reunion photo…
04-01-2004 Hilton, Jacksonville Beach, Jenny, Bill, Arthur, Janine, Uncle Greg, Grandpa Whitey, Phillip, Rosemarys baby, Aunt Josaphine, Rover.jpg.
Now do that to the other 50,000. See the problem? If I did this, I would not have anytime to live, or take the pictures for my library.
3 – Separate Index
Many great programs are making strides at image organization. The best I’ve tried thus far is Picasa (owned by Google). It creates a timeline of your photos, allowing you to “tag” the photos with relevant information. The problem with this, is that the tags are not part of the source image, rather they are relational information in the Picasa software. So if you rename the file in Windows Explorer for example, the linked tags in Picasa are orphaned. This issue is in no way isolated to Picasa, nearly every image management program has this fault.
So How Do I Organize my Images?
At the beginning of each year, I build that years “containers” for each month (e.g. 2009-01 through 2009-12). Naming the folders Year-Month-Day, forces Windows to sort them chronologically. As I take photos throughout the year, I put them in their respective folder. If there are special events during the year they get their own folders, like “2009-03-12 Jim’s Retirement Party”. Doing this makes finding images relatively easy, and it keeps everything sorted.
As for all of my scanned images, I just put them in the best guess folders, if the date is not known. Because there is no “index”, you are free to move things around all you want.
The one flaw in this system is finding images of a particular person. doing so requires manual scans of the images. But the date system virtually eliminates duplicates, so it speeds up the search in some aspects. I’ve thought about building a way to tag the images, but I know as soon as I get close to being done, someone will have built what I want…
What I Want
We read in the news all of the time about the “face recognition” software used to fight terrorism. When will this power and speed be offered to consumers. Imagine software that would allow you to choose “selector” images, which are good facial shots of your friends and family. The software would “facemap” these images and “learn” what the person looks like. You would be able to upload multiple images of the same person, so the software would learn each person’s facemap at different ages.
The software would then index your entire image collection, searching for matches to that face map, and then added it to “extra” EXIF information on the image, in the forms of metatags. Randomly, the software would ask you to verify its findings, which it would use to become even smarter.
Locations could also work this way. That picture of Aunt Martha standing in front of Alamo, would automatically be tagged with “aunt martha”, “Alamo”, and be geotagged San Antonio.
So until that awesome software comes out that I really want, I’ll be sticking to the yyyy-mm-dd method of organization.
Do you have a tried and true way that you use? If so, explain it in a comment.
Image: QT1p [via Flickr]