Case study - image manipulation and production
Photoshop: Compositing, Masking & Visual QA
This case study was developed to demonstrate my image production skills in the context of AI workflows — an area that intersects directly with my background in graphic design and visual production across the Adobe ecosystem.
01 - Overview
Overview
My design background spans graphic production, web design, UX/UI, and visual systems work across tools like Illustrator, Figma, and Lightroom. This case study documents three focused personal Photoshop projects — masking, compositing, and visual quality review — that reflect where those skills intersect with image editing and AI-assisted workflows.
I've been working hands-on with Adobe Firefly as part of exploring how AI generation fits into a professional production process, and these pieces reflect that intersection — understanding not just how to manipulate images, but how to evaluate and quality-check them too.
Masking, compositing, and quality review represent the core of what makes image production work at any scale — getting clean edges, making scenes feel believable, and having a sharp enough eye to catch what's off. These felt like the right place to start.
02 - Piece One
Precision Masking & Complex Selections
Using Select Subject, Refine Edge, and manual mask painting to achieve clean, production-ready cutouts — including fine detail like hair strands and semi-transparent edges.
Getting a clean mask is harder than it looks — especially around hair and fine edges. This piece is about showing the difference between a quick auto-selection and a result that's actually production-ready.
Original — subject on original background
Edge detail at 200% zoom | Raw auto selection:
After Refine Edge pass:
Final manual cleanup:
After — clean mask with refined hair edges
03 - Piece Two
Image Compositing & Lighting Integration
Combining multiple source images into a single believable scene — matching color temperature, light direction, and shadow placement using adjustment layers, blend modes, and manual dodging and burning.
A composite only works if it's believable. The challenge here was making two images that were shot differently feel like they belong in the same world — matching the light, the color, and the way shadows fall.
Source images and layer structure | Background — [describe]:
Subject — [describe]:
Layer organization
Final composite — [brief description of what you combined]
04 - Piece Three
Visual QA — Artifact Identification & Correction
Identifying and documenting common image quality issues — including edge bleeding, color banding, halo effects, and unnatural compositing — and applying targeted corrections to bring images to production standard.
Having a good eye isn't just about making things look nice — it's about knowing exactly what's wrong and being able to name it. This piece is about that: spotting the specific issues, documenting them clearly, and showing the fix.
BEFORE — Artifact Version
Issues Identified & Resolved:
AFTER — Corrected Version
Edge bleeding — Color from background spilling onto subject edges along the left shoulder. Corrected with targeted masking and hue/saturation adjustment.
Halo effect — Bright fringe visible around hair due to aggressive background removal. Resolved with decontaminate colors and manual mask refinement.
Color banding — Visible tonal steps in the gradient sky area. Corrected with noise addition and gradient smoothing.
Resolution inconsistency — Subject and background were at mismatched resolutions. Resampled to consistent 300dpi output.