(Early Access) Assisted Culling (Lightroom Ecosystem - Desktop)

I'd like to thank the addition of AI culling (keep working on it) and the improvements in metadata handling (better string management and visualization).

IDEA:

Now it would be great to have AI editing like imagen.ai.

WHY:

1 - It would be great to have it integrated in my favourite and most used tool

2 - Imagen AI works by pretending a lot of different pictures edited the same way while Lr could learn from many single editings and learn and evolve better, with less pain for the photographer

3 - It would improve by order of magnitude the time I spend on editing (mainly weddings, the most time consuming case and where the largest number of pics are delivered)

I have just started using the assisted cull feature. One place where I am having a problem is in a senior photo shoot where the senior wanted some photos with their shades on. The cull feature accepted the focus for the subject and the eyes but rejected it as his eyes being closed. There should be a way to override the red Xs. Like be able to reject it and send you feedback to make the system smarter.

Hi,

I believe this feature is still new and need improvements, I think one feature that everyone would want is the ability to select the best shot from a series of similar shots and this will save a lot of time.

Regards,

Wilson

With all due respect for the team's effort, I find this culling feature of little to no value for my work. I have trained eyes, a clear vision, and direct my models deliberately. I tested the feature seriously in my field: portraiture as an artist — theatrical productions under difficult lighting, fantasy/DOP projects, boudoir, and commercial shoots with real people. I do not shoot weddings, sports, street photography, or documentary work. I create intentionally; I am not a novice and I do not shoot thousands of images to choose one.

My suggestion to Adobe Product Management (Lisa Ngo and Kwamina Arthur): make this an optional paid feature. I disabled automated culling because it offered me no benefit, and I have never missed such a function.

please make a version by which it will automatically choose between similar images. stacking is not doing it yet.

SO excited for the new feature! What would be really helpful it to be also to do assited culling on PARTS of a whole gallery. With bad wifi or low ram it can take a long time to run and to be able to use it in the places it would save me the most time would be really helpful. Also a pause, stop and save button would be really helpful.

I’m thrilled to have this feature now! I’ve been hoping for it.

I’d like to give some feedback on this feature. As a multipurpose photographer who also uses my photos for graphic design, I often need to find specific car photos. However, I don’t have enough time to organize over 200 race cars into separate files. It would be great if I could organize photos by their numbers displayed on the side of the car, or if there’s a feature that allows me to highlight specific areas and then create stacks from the search results.

Thank you to everyone who works hard to make this program user-friendly! 

Love the idea of making culling quicker! However upon first use, I ran into an issue with the batch action option. I selected a specific album and began culling (eyes open filter didnt' seem to work very well and missed some obviously closed eyes in single person portraits, but I didn't see an option to mark it as incorrect, just to mark it as a reject). I'd gotten to a point I was happy with, then selected batch option to rate all the selects as having a ranking of 5 stars (this is how I organize in-album). When I then navigated to the album later, it had not marked the selects as 5 stars and I had to start over again, just doubling my culling time (sobbing!). Love this concept, but deeeeefinitely keep tweaking it. Thanks!

Need an obvious way to stop the culling process.. 

As a high-volume sports shooter, I’m genuinely disappointed that Adobe engineers could not implement a simple AI-assisted "ball-in-frame" detection feature and be the first to do so. 

When you look at the photographers who cull the most images, two groups stand out immediately, wedding and sports photographers. Seeing the below comment about culling 300 images is laughable to those of us who regularly cull 5000+ after a normal Tuesday.  Sports photographers, in particular, rely on burst shooting to capture the split second where the athlete, the motion, and the ball align perfectly.

The fact Adobe engineers hadn't figured or didn't prioritize an AI scan of images for a ball is wild to me.  The first major platform to that party, whether Aftershoot, PhotoMechanic, Imagen, Adobe, etc.....will dominate sports photography photo processing for at least the foresable future. Adobe might even win some customers they've lost form their constant price increases.  

Here are a few things for Adobe engineers to consider as to why to prioritze this feature moving forward:

• A modern sports camera can shoot 30–40 frames per second (not including the A9iii global shutter of 100+)

• Out of a 10-second play, you may have hundreds of frames with only a handful that actually include the ball.

• Manually culling those frames is extremely time-consuming and easily automated with AI because surely AI knows how to spot a ball right? 

This is not a complex ask. Adobe already has object detection in Lightroom and Photoshop for people, skies, backgrounds and now things like reflections. Extending that detection to identify “ball-in-frame” specifcally in this new culling feature would be an immense productivity upgrade. 

Frankly, it’s surprising Adobe hasn’t already prioritized this, especially given the recent price increases. Features like this are exactly what justify the continued loyalty of high-volume professionals. 

Adobe engineers,  add a “Ball-in-Frame” AI filter to this culling feature in the next update please! It would not only save countless hours for sports photographers worldwide but also reaffirm Adobe’s commitment to innovation and responsiveness to the real-world needs of its users.