P: Culling tools to pick sharpest image

[Bob Trlin]

These days with high bust rates used more often, culling has become quite a chore.  I notice that in OM WorkSpace, there is a facility that selects a number of the sharpest images in a burst series of like images.  If this can select say 5 of the shapest images in a burst of 20, that reduces the culling workload by 75%.  I would think this a very worthwhile feature.

[naestyoh]

I'd like to see Adobe use their new AI chops to help filter images. 

• AI or user would create stacks. The statistics would be applied to that stack.
  • AI: build stacks based on a combination of time-between-shots and feature matching
  • User: Select images, create stack
• Use a simple edge-detect laplacian kernal convolution and calculate variance
• Rank the images based on deviation from an ideal value: 5-star image if it's within some determined acceptance. Rank the remaining based on distance from that image. If they are all mostly sharp images, you may get one 5-star and 4-stars for the rest. If you were spraying-and-praying, maybe one shot is 4-star and the rest are 1-star. Use a badge to show the sharpness and deviation. Use this info for HDR or other stack-functions where you want the set to be above a threshold. 

^^ 400% enlargement shows the differences not easily seen at 4k/5k monitor res. 

[johnrellis]

Moderators, @Rikk Flohr: Photography, please merge with this very similar idea:

https://community.adobe.com/t5/lightroom-classic-ideas/p-culling-tools-to-pick-sharpest-image/idi-p/...

[PhotoVid]

Please add AI Culling features to Lightroom Classic to automatically start rate or color code images after Lightroom has analyzed and a selected group of photos. Lightroom should be able to detect duplicates and select the best of the bunch based on sharpness, smile detect, rule of thirds/composition. Additionally, culling should also automatically mark blurry photos (out of focus/ unintended motion blur), photos too far out of dynamic range parameters (too dark or to bright to be fixed), closed eyes, etc. 

Bad photos should not be deleted but simply categorized via star ratings, color coding, and or keyword tagging so that we can filter the results we want to use or recategorize photos. 

This functionality is currently presented in AfterShoot, Narrative Select, PostPro Wand, FilterPixel, Optyx, DopeAI, (PhotoMechanic), etc. We should not have to leave lightroom or be subjent to these additional tools for this basic functionality since culling is CORE to what LR was created for AND because LR now has the ability to analyze subjects find face details using the subject select tool. 

[agustb33249581]

Hi, I'm a student in digital media and not unsurprisingly I take a lot of bad or unnecessary photos. Sometimes I take a thousand photos in one day while capturing an event for instance.

When I import into Lightroom Classic there are very few options. I cannot even auto rotate the photos on import as in most other programs.

The problem with having a 1000 images and a deadline is that the time it takes me to flag or rate my collection increasingly eats at my deadline. I often spend hours just going through collections without doing any actual work.

One simple thing would be for Adobe Lightroom to have an automatic ranking Ai helper which would choose the best and worst photos.

Since photos are digital it should be very simple to automatically sort out over exposed, dark or blurry photos from others which are within the most common criteria.

• Does anyone know if Adobe Lightroom has any such options?
• If not I have the student Adobe Creative Cloud package and are there any other Adobe products that can rank photos automatically?

Hope someone can advise me,

Thank you,

Agust

[johnrellis]

Moderators, @Rikk Flohr: Photography, please merge with this existing Idea:

https://community.adobe.com/t5/lightroom-classic-ideas/p-culling-tools-to-pick-sharpest-image/idi-p/...

and change the title of the merged thread to "AI-based culling", which better reflects the content of the existing posts in that thread.