Removing Glare from Glasses: A Real Before/After Test

My LinkedIn photo has been the same for four years. Not because I'm lazy — because every time I tried to take a new one, my glasses turned into mirrors. Streetlights, windows, my own desk lamp. The reflection spots moved with my head but never disappeared. I just kept using the old photo.

Then I spent an afternoon testing whether AI photo editors can actually fix this. The results surprised me.

The Problem Setup

I collected forty-seven photos from my phone and camera that had glass glare issues. Some from a work conference where the venue had huge windows. Some from a dinner where the restaurant had those Edison bulb chandeliers that look great but create nightmare lighting. A few from a rainy day where wet pavement reflected everything.

The glare types varied — some was subtle white spots on the lenses, some was full window reflections showing the room behind me, some was the dreaded "complete stranger visible in my glasses" situation that happens when someone stands behind the camera.

I tested four AI tools: AIPGEN, Adobe Express, Snapseed, and an online tool someone's friend recommended on a photography forum.

AIPGEN — The Object Removal Approach

AIPGEN doesn't have a dedicated "glare removal" button. What it has is an object removal tool that works on anything in the frame. I selected the glare spot as the object to remove and let the AI heal it.

For subtle glare — small white spots from point light sources — this worked surprisingly well. The AI sampled the surrounding lens area and filled in the spot with what it predicted should be there. In most cases the result was accurate enough that I couldn't find the original glare location without close inspection.

For complex reflections — windows, rooms, other people — the results were more mixed. The AI removed the reflection but sometimes introduced soft artifacts where the fill looked slightly "guessed." In a LinkedIn profile photo at normal viewing size, this is fine. In a print or cropped detail, it's noticeable.

The before/after slider in AIPGEN made it easy to compare and decide whether the result was good enough to use or needed another attempt. Most photos took two to three rounds of object selection to get a clean result.

Adobe Express — The Automated Approach

Adobe Express has a "remove reflection" feature that's marketed specifically for glasses photos. It works by detecting the glasses and attempting to remove light reflections automatically.

The automatic detection worked well — it found the glasses in almost every photo and applied the removal without me having to select anything. The results on subtle glare were comparable to AIPGEN's manual approach. For complex reflections, the automated system sometimes over-corrected and left the lenses looking unnaturally clear — like someone had removed the glass entirely and left a wireframe.

The subscription cost is harder to justify for occasional use, but if you're regularly photographing people with glasses, the automatic detection alone might be worth it.

Snapseed — The Manual Approach

Snapseed doesn't have AI-powered object removal. What it has is a healing tool that works similar to Photoshop's content-aware fill. You brush over the area you want to fix and it samples from surrounding pixels.

This required much more manual effort. I'd brush over the glare spot, see the result, and either accept it or undo and try again with a different brush size. For simple glare, it worked adequately. For complex reflections, I spent too much time trying to get a result that didn't look patched.

The advantage is Snapseed is free. The disadvantage is it requires actual skill and patience. I have neither in large quantities.

The Forum Tool — A Cautionary Tale

I tried an online tool someone's friend recommended on a photography forum. It claimed to use AI specifically trained on glasses reflections. The results were terrible. It removed the glare and replaced it with smeared color artifacts that looked worse than the original problem. Two photos I tested this on are now permanently corrupted in my mind as "the ones where I learned to read forum recommendations more carefully."

What Actually Works

After testing forty-seven photos, here's the honest summary:

AI tools can remove glasses glare. They work best on subtle to moderate glare from point light sources. They struggle with complex reflections — windows, rooms, detailed scenes — but can still produce usable results with patience and iteration.

The key insight is that "removing glare" is really "replacing the glare area with predicted content." The AI is guessing what should be behind the glare. The accuracy of that guess depends on how much surrounding context exists to sample from. A small glare spot on a consistent lens color has lots of context. A large window reflection covering most of the lens has less.

Results are good enough for social media photos. They're not good enough for professional headshots where someone is paying you to deliver perfect files. For professional work, you still need to control the lighting environment when you take the photo.

My Actual Workflow

After this testing, here's what I do now:

• Try to get the photo right in camera — position lights to minimize glare, have the subject tilt their head slightly to change reflection angle
• If the glare is subtle, use AIPGEN object removal and be done in one round
• If the glare is complex, spend three rounds of iteration and accept that at social media size the result is fine even if it's not technically perfect
• Never use the forum tool

I updated my LinkedIn photo. The glare is gone. I look like I know what I'm doing, which is the actual goal.

One Thing to Know Before You Start

The best glare removal happens before you take the photo. Position your light sources 45 degrees to the side and slightly above the subject — this angle reflects light away from the camera and toward the floor. Have the subject remove their glasses briefly while you check the lighting, then put them back on for the actual shot.

AI tools fix what you couldn't fix in camera. They work best as a safety net, not a primary workflow. If you're regularly photographing people with glasses, learn the lighting positioning first. You'll need the AI less often.