appssemble
ServicesBlogCase StudiesAboutContact
Case Studies/HDR Plus+

DSLR-quality photos from your phone

Capture bursts of up to 8 RAW photos, align and merge them with custom TensorFlow models running entirely on-device. Full manual control over ISO, exposure, and white balance. AI auto mode for quick shots.

View on App Store
RAW FRAME 8/8+2 EVISO 64ML MERGEalign → denoise → tonemapHDR OUTPUTQUALITY8 RAWframes per shotON-DEVICE MLTF LiteMetalPIPELINEcapturealigndenoisemergetonemap
8 RAWframes per burst
5.0App Store rating
On-deviceML inference
What it does

Pro camera, pocket-sized

01

RAW Burst Capture

Capture up to 8 RAW frames in a single burst. The camera pipeline handles alignment, exposure bracketing, and buffer management. Every frame is a full-resolution RAW file ready for ML processing.

RAW captureBurst mode8 frames
02

ML Merge & Tone Mapping

Custom TensorFlow models align and merge the burst stack on-device. Noise reduction, ghost removal, and HDR tone mapping happen in one pass. The result looks like a DSLR shot.

TensorFlowOn-device MLTone mapping
03

Manual Controls

Full manual control over ISO, exposure, white balance, and focus. EV presets for quick adjustments. Shoot in auto mode or take complete control — the interface adapts to your skill level.

ISOExposureWhite balance
04

Professional Editing

Edit merged photos with pro-level tools: color grading, vibrance, skin smoothing, depth control, de-ghosting. Apply presets or create your own. Every edit is non-destructive.

Color gradingPresetsNon-destructive
05

Import & Process

Import photos or HDR brackets from any camera or app. Run them through the same ML merge pipeline. Get DSLR-quality results from any source, not just the in-app camera.

ImportExternal bracketsUniversal processing
Under the hood

What powers it

Swift

Native iOS app built for performance-critical camera and image processing pipelines.

TensorFlow Lite

On-device ML models for frame alignment, noise reduction, and HDR merging.

Metal

GPU-accelerated image processing for real-time preview and editing operations.

Core Image

Apple's image processing framework for filters, color adjustments, and compositing.

AVFoundation

Low-level camera control for RAW capture, burst mode, and exposure bracketing.

Figma

Design system for the camera interface, editing tools, and preset management.

Process
Agile2-week sprintsDaily standups
Technologies
iOS (Swift)ML (Python)TensorFlowFigma
Takeaways

What we learned

01

Photography is a deep domain

Learning pro-level photography was one of the best parts of this project. Our team now speaks fluently about exposure bracketing, tone mapping, and color science. Domain expertise made better software.

02

ML on-device is the future

Building custom TensorFlow models that run entirely on the phone was the most challenging and rewarding part. No server round-trips, no latency, no privacy concerns. The phone does everything.

03

ML applies everywhere

This project showed us that machine learning can solve problems we thought were impossible on mobile. It pushed our AI team further and opened new capabilities we now bring to every project.

Case studies

More work

10M+events / day
Grovs
Open SourceAttributionDeep Linking
→
500K+app sessions
Semaphr
SaaSiOS SDKAndroid SDK
→
<30sdraft to e-Factura
Incasez
Invoicinge-FacturaSaaS
→
100%offline ML
Dezigner
iOSARML
→
Let's talk about your project
[email protected]
CompanyServicesCase StudiesBlogContact
Offices
New York1740 Broadway, 15th Floor, 10019
LondonKemp House, 160 City Road, EC1V 2NX
Cluj-NapocaBlvd. 21 Decembrie 1989, 95-97
SocialLinkedInGitHub
© 2026 appssemble. All rights reserved.
Privacy PolicyCookie PolicyJobsGlossary