I’ve got hundreds of thousands of photos sitting on my Synology NAS, and the fact of the matter is that a significant portion of them are duplicates that I haven’t gotten around to yet. Some are exact copies, while others might be compressed versions shared by others. There are even versions created during photo editing sessions. Finding them manually is just not practical. I simply don’t have the time for it.
It’s my search for a deduplication tool that led me to the annoyingly named dedupPictures. This app, as cryptic as it sounds, can talk directly to my Synology NAS instead of downloading every photo to my Mac, using the thumbnails that Synology already generates. It then compares them locally and gives me an easy interface where I can decide what stays and what gets deleted. It’s helped me recover hundreds of gigabytes of space from my NAS.
I’m torn between these 5 user-friendly NAS image hosting tools
Spoilt for choice
My Mac does the heavy lifting
It doesn’t need to download the original photos
The big problem with deduplication tools tends to be the way that they access a NAS. If I mount my Synology over SMB, most tools need to pull down full-size images over the network to determine whether the two photos are the same. That’s particularly inefficient when I have RAW files or large JPEGs. Downloading terabytes of data is just inefficient.
DedupPictures takes a better approach when deployed with a Synology NAS. It connects to Synology DSM’s FileStation API and pulls down the NAS-generated thumbnails instead. Those thumbnails are obviously an order of magnitude smaller in size. Compared to the full images, you end up saving a huge amount of data, and the app works off these thumbnails to inspect the library without moving the full-resolution images across the network.
In fact, this is the biggest reason why I opted for the app. My photo library stays on the NAS, while my much faster Mac does the comparison work. I can point it at the relevant folder, let it scan the files, and get results without filling up my Mac’s storage. And let me be real, I don’t even have enough storage to pull it off on my Mac.
I get to choose what gets deleted
Perceptual hashing catches different versions
Close
The other aspect that stands out is that the app doesn’t rely just on exact file matching. It uses a perceptual hash that is designed to recognize photographs that look the same even when the files are not identical. It’s this hashing algorithm that catches cases when an image has been resized, recompressed, or sent through WhatsApp and saved again.
This was particularly important to me as my photo library, stretching across hundreds of thousands of images, has been compiled over decades. Invariably, there are duplicate images, versions, compressed copies, and even files with differing file names. Just file name-based deduplication simply wouldn’t work for me. For the file system, they are separate photos. But technically, they might be the same image.
DedupPictures maintains a cache of its hashes so that once you’ve scanned the library, you don’t need to repeat the analysis with subsequent runs. It makes it particularly useful for longer projects where you want to maintain your library under control as new photos are added in.
Of course, finding duplicates is just one part of the problem. As good as the software is, I don’t trust it to decide exactly which photos to delete. No algorithm is perfect, and an accidentally deleted photo is not acceptable.
Instead, dedupPictures gives me a local web interface where duplicate groups are displayed together. While it automatically selects a file to keep, favoring the larger of two or more files, you have the option of changing the selection. You can make all of these selections, hit delete, and let it handle the rest.
You’ll also find a semantic mode that solves the problem of retakes. Perceptual hashing is great at recognizing copies of the same image, but it can’t identify images that are close enough but slightly different. Instead, the app, when in semantic mode, uses a local CLIP vision model to compare what’s depicted in the image.
This can be useful when cleaning up photos taken in burst mode and repeated shots. Often enough, while out and about, I’ll capture multiple photos of the same subject to make sure I capture the frame. But I don’t really need all these files. I only need the one or two images that captured the setting. The CLIP model can help identify all these images. It previews and reviews these images, and serves up the web server for you to choose the images to keep or delete.
The easiest way to keep my NAS photo library clean
Removing duplicate photos is a big part of maintaining a clean photo library. But more than that, in this economy, anything that saves me having to buy another hard drive by releasing wasted space is an essential tool.
DedupPictures handles my mess of a photo library elegantly by removing duplicate photos, copies, resized images, and much more. And it does this without having to pull down my entire image library from my NAS. That alone makes it a worthwhile addition to keeping my photo library clean.
dedupPictures : This photo deduplication tool finds duplicate and visually similar images to help you review and remove extra copies.