Screening, Categorizing and Retrieving
This will wrap up the Image File Organization step which so far has considered –
- General Overview
- File backup
- File folder options and file naming suggestions (done during image download)
This final section will look at adding data to each image file in order to facilitate its retrieval in the future. We’ll consider –
- Image screening
- Image labeling
- Image categorization
by showing steps that I take in this process. The examples are based on the digital image management software that I use, IMatch, but other programs exist as well.
As always, click on any image to enlarge.
This image picks up where we left off in the previous post – a new folder, 100925 frog, has been created by the download program and the image files from the memory card were named per my naming specs and added to the new folder.
Preliminary screening – Now that the files are downloaded, what’s next? For me, it’s time to do a preliminary screening primarily to determine which files I may want to delete immediately. If you look at the thumbnails you’ll see that I tend to make several nearly identical captures (often varying just the depth of field) and during this screening step decide which to keep – sometimes all, sometimes a few, sometime none. No right answer here – dealer’s choice; it depends on you.
There are lots of ways to do this screening step. The basic requirement is for a browser that will allow you to label or rate your images on the fly. The faster the browser the better and the more labeling/rating options the better. Instead of flagging keepers you might simply delete the unsatisfactory images on the spot – but I don’t recommend that unless you never change your mind about anything.
Shown below is the “browser” built into my digital image management software, IMatch. (I use IMatch for every aspect of step #3 and even call the editing and printing programs from within IMatch – it’s my command and control center). Back to the browser shown below – note the little yellow symbols in the upper right corner of some thumbnails. IMatch allows me to scroll through the images and mark them with either
- A bookmark (the yellow symbol)
- Any one of five labels (color coded bar below the thumbnail)
- 1-5 stars
- or – any combination of the above, including none
On my first screening pass, I simply bookmark the ones that I think I’ll keep as shown here.
After bookmarking the images to be Kept, I return to IMatch’s main screen (hit Esc while in the browser) as shown below. This is the next to last chance I give myself to review all of the images before actually deleting some.
Next, using image filter options shown in the drop down menu illustrated in the next image, I filter out the “keepers” so that I can take a final look at the images planned for deletion. In the image below I’ve selected the filter “Show No Bookmarks”. New filters can be created and added to this list to suit your needs. Further, the filters can be combined – for example I could have selected No Bookmarks AND JPEG Only (which would show zero images since at this stage they are all RAW image files – no JPEG’s).
After this final check, I did a Select All and hit Delete – and confirmed the “Are you sure…?”.
After deleting the above images, I cleared the filter (screen was blank at this point since the filter was set to no bookmarks and all the non-bookmarked files were deleted) and got the following. Another select all followed by clearing the bookmarks prepares me for my second screening pass.
Second Screening – After deleting the non-keepers from the initial screening and removing the bookmarks from the keepers, I’m ready for a 2nd pass. This time I’m looking for those images that I might wish to process. My method for doing this is similar to what was just shown for selecting keepers as shown below – browse through the keepers and bookmark potential candidates for further processing/editing.
Once again, returning to the main page after the browse-bookmark step, we now see the selected candidates. As opposed to using a filter to show those not bookmarked so that I can delete them (as done in the preliminary screening – these I want to keep, but just not process now), this time I used a filter to show only those images with a bookmark as seen in the image following this one.
Below we see only the images identified as potential processing candidates by our second screening pass. Because of my penchant for capturing several similar images, at this point I often make a 3rd pass.
3rd and Final Screening Pass – The image below shows us back at the browser following the screen shot above. Note that the browser knows that a filter was used to show only bookmarked images and, thus, shows us only those images and not everything in the current download folder. In this pass I’m looking for the best of the best (not that any of these images is anything special). I flag this next group by adding a star to the selected files (it appears below the image in a gray bar which would be one of 5 selected colors, instead of gray, had I chosen to add a label in addition to or instead of a 1-5 star rating – lots of options available).
Shown below is the final result of the screening process (as I do it; everyone has their own way & I doubt if many follow mine). I have the filter drop down menu shown to illustrate that I’m focused in on my final set, keepers that are candidates for processing & editing, a total of six out of our initial download of 27 images.
It took far, far longer to describe the process for getting to this point than to actually do it. In actuality, it was only 2-3 minutes.
Screening Fine Tuning – At this point four of the six finalists are still nearly the same image. A good image viewer will allow a closer side-by-side comparison. IMatch will allow up to four images to be compared side-by-side as shown below. There are additional side-by-side options such as histograms, flashing highlight/shadows, etc. that can be used for a more detailed comparison. In this case I’m using the side-by-side to select one or two of these four. You can see via the drop down box that side-by-side view allows me to set ratings, labels and bookmarks, just as the browser did. In this case I’m changing a rating from 1 to 2 stars.
After changing two of the four frog images from 1 to 2 stars, the screen shot below shows the results of filtering for all images with a rating greater than zero. I normally start processing with the highest rated and work my way down. In this case I eventually (not at this stage in the image organization step) processed four images – the two non-frogs and the two 2-star frogs – from the original 27.
At this point, about 5 minutes into Workflow Step #3, we’ve
- Created and named a folder for our current set of images
- Downloaded and named the image files
- Screened the files
- Identified keepers
- Flagged a subset of the keepers for future processing
That’s all necessary and good, but except for potentially useful file folder structure and naming coupled with helpful image file naming we haven’t done anything organization-wise to help us with image retrieval a month or a year from now. That remains to be done (and could have been done prior to the screening – your choice). We need to add data to each image that will allow them to be identified by something other than date or file name – specifically to pick them out of a crowd by their subject matter, shooting location, genre, etc. Let’s take a look a how one might do that.
Ideally for each image we’d like to know some or all of the following to facilitate future retrievals –
- when it was taken – usually handled for us by the camera
- where it was taken – potentially handledby the camera if it’s GPS capable)
- who/what the subject was – the who is potentially handled if the camera/software handles facial recognition
- genre or style (such as macro, infrared, sunrise, abstract – one or more may apply)
So far we’ve taken care of the when, but none of the rest. Options abound – much depends on what your software will support. Generically we can handle the missing three items above by adding things like –
- Tags (these would include bookmarks, ratings and labels as shown earlier, but they are non-specific)
- IPTC/XMP metadata added in camera (or afterwards) for camera data including lens type, shooting mode, etc.
My personal preference is to use categories created in IMatch. This isn’t the way to go if you are doing things like stock photo where the images are sent to someone else. The categories are maintained in IMatch and not physically added to the image file or to a sidecar file (a separate file that is associated with the image file). As such, my category data would not be of any use to someone not using IMatch AND having a copy of my database’s category data. If the image retrieval data is important to others, besides yourself, you’ll need to use the IPTC/XMP options in programs such as Photoshop that are compatible with those standards.
Categories – This is how I use categories. Extrapolate this to other tagging/labeling/key-wording options that suit your needs in order to accomplish the same result.
Programs that offer a categorization option will allow you to create your own. This is highly personalized and dependent on what & how you shoot. I am mostly an outdoor photographer. The category outline shown below supports my What/Where/Genre data needs (and even adds a bit of When to the mix to complement the when data already embedded within my folder and file name structure – you’ll note there is a Season category which allows me to retrieve images by season as one of more of a search criteria’s components). The Criteria shown are editable – I can change/add/delete to suit my needs as time goes by.
The three main category areas in my system are –
- What – covered by Fauna, Flora and Landscape
- Where – covered by Places
- Genre/Style – Covered by Style (more on this one later)
As shown below, the Flowers portion of the Flora category is expanded and within that set the Flower category called Bulb is also expanded. If I just shot 25 tulip images, I simple select them all on IMatch’s main screen shown above, select Category Assignments, click on the box next to Tulips and it’s done. I could also open the Season category folder and assign the spring category as well. Similarly, if some of the tulip images were macros I could select that subset, open the Style category, and select Macro – and on and on and on….
But back to today’s download – water garden plants and a frog. As shown below, I’ve selected all of today’s images (not just the top 6). Every image has water plants in it even if it’s not the focus of the image. You can see that I’ve selected the water sub-category of Flowers for every image. Also, you can see that the water category of fauna is highlighted and will be selected – but just for the images that include the frog. As a preview of things to come, later when I need to retrieve images I can ask to see water based flora BUT with NO fauna and retrieve only the “frogless” subset of today’s images (or any other combination that you can imagine).
OK – continuing from above, I’ve selected only the images that include the frog and clicked on the water based Fauna category.
One final bit of categorization. I’ve placed all images used in this Photo Improvement blog into a Photo Course Blog category. Although this batch were done at download, others used in the blog predated the blog – no problems as image categories can be updated and modified at any time. Lastly I’ve put the images into HH under Places (where) – where I live.
I haven’t added a Style category for the frog images. I could but experience tells me that I’ve included all that I’ll need (and if I’m wrong it can always be added later). However, the screen shot shown below illustrates the types of styles that I typically use. Imagine wanting to retrieve a sunrise shot in Maine (I have 1000’s of sunrises and 1000’s of Maine images). Being able to ask not only for images made in Maine but to add AND including a sunrise is powerful (and simple)
This wraps up my brief example of the types of activities go into image organization to facilitate future image search and retrieval. This may seem like a long tedious process but if you have the right software it’s not – everything that I described for a typical day’s shoot takes about 10 minutes (of course if you shot a 1000+ image event it will take longer, but something this side of 100 certainly shouldn’t take much more than 10-15 minutes – time well spent!).
Previously I mentioned that my image management program, IMatch, is used for everything including calling other programs such as editing, noise reduction, printing, etc. The screen shot below shows the list of programs that I can call from IMatch (editable, of course). Just select the image or images (highlighted in the example) to be processed and click on the desired program in your “User Tools” and the selected image(s) appear in the selected program ready to go.
A few retrieval examples and we’ll leave Workflow Step #3 – finally!
The following retrieval screen shots show the IMatch category view (all earlier images were of the database view; they differ primarily in the contents of the left window). You’ll need to click and enlarge these images to follow.
Example 1 – Here is a request for all images categorized as Water Fauna. We see the frog images just completed plus a couple for turtles already in the database. Note that to do this I simply clicked on the Water entry under Fauna and waiting 2-3 seconds. There are 4 windows here – note that the bottom right window lists the categories assigned to the highlighted thumbnail (just what we assigned above).
Example 2 – What if we wanted to see images of Water Fauna but only if there were Water Flora in the image as well. Take a look at the bottom left window. It says to show only images that have both. I did this simply by clicking and dragging the desired categories from the upper left to the lower left window. (To repeat – I just dragged & dropped; all of the typing shown in the lower left window resulted from my D&D)
Example 3 – How about if instead of Water Flora AND Water Fauna we wanted Water Flora and Insect Fauna. This image shows what we get by repeating the steps from example 2 but substituting Insect.
Example 4 – Following on with example 3, supposed we wanted only color images (that is, remove all B&W). Simple – drag the Style name B&W and drop it into the “NOT” column. As we see in the lower left window this gives us what we want – (“Flora.Flowers.Water” AND “Fauna.Insect”) NOT “Style.BW” – all images containing both insects AND water flora (those are lotus) but NOT B&W,
Example 5 – Of course, the opposite of #4 is equally easy – just drag the B&W style to the AND column instead of the NOT column.
A few observations before stopping. The examples are simplified for illustration of the types of things that are possible. Once a retrieval is complete it can be refined. For example, the same image filter is available for category results as was used in our earlier database screening – we could ask of the retrieved images to see only those with specific combinations of filters such as only Red labels, rated 3 or higher, with no bookmarks and just RAW versions – and on and on.
Most of my retrievals rely on the Places and Styles categories with the “what category” options such a Fauna, Flora, etc. coming in third.
That’s it. Thanks for sticking with this to the end – you are probably either the first or second. 😉