Vehicle imaging platform
Studio-standard car photos, captured anywhere
Quality stops depending on who happens to be checking.
What we did CodeLeap built iOS and Android apps that guide each shot, a custom computer vision model that locates the vehicle and its features, automated background replacement and quality control, and 360-degree turntable spins, on infrastructure handling millions of images a year at a cost per image that keeps the product commercially viable.
Studio car photos, anywhere. Our client works in automotive imaging, producing vehicle photography at a volume where manual checking cannot keep up. Inconsistent framing, stray backgrounds and missed angles were reaching live listings, because no team can review that many images by eye. Over a multi-year engagement, CodeLeap built the whole pipeline: mobile apps that guide each shot, computer vision that cleans and standardises every image, and automated checks that catch problems before they go live.
Product features
-
Guided capture apps
Photographers capture each vehicle through iOS and Android apps that guide every shot on angle, framing and coverage. Solving consistency at the point of capture removes an entire category of problem, because an image shot correctly never needs fixing or rejecting later.
-
Custom computer vision
A model trained specifically for this application identifies vehicle positioning within each image and locates individual features such as wheels, windows and steering wheel. Every downstream step depends on the system knowing precisely what it is looking at.
-
Automated background replacement
Computer vision separates each vehicle from its surroundings and drops in a clean, professional backdrop. A retailer can present every vehicle at showroom standard regardless of where the photograph was actually taken.
-
Automated quality control
Every image is checked automatically for framing, missing angles, blur and other defects against a defined standard. Problem shots are flagged for attention instead of quietly reaching a listing, so human attention goes to the exceptions.
-
360-degree turntable spins
Approved images are assembled automatically into interactive 360-degree spins, served through the platform's web media player. Buyers get a richer and more consistent view of each vehicle, with no manual editing step.
-
Processing infrastructure at scale
The infrastructure handles millions of images a year and scales through peak demand without degradation. For a platform whose economics depend on cost per image, that balance was as much a requirement as the visual quality.
Challenge
The client had built a substantial business in vehicle imaging and wanted to raise the consistency of the images already flowing through their systems. Volume was what made this difficult. With millions of images a year, inconsistent framing, stray backgrounds and missed angles were slipping through to live listings, because no team could review that quantity by hand.
Everything therefore had to be automated end to end, which set a demanding technical problem: taking raw photographs shot in car parks and forecourts by people of varying skill, applying computer vision to identify each vehicle and its individual features, standardising and enhancing every image, inspecting the results against a quality standard, and assembling interactive spins, all without human intervention. The economics mattered as much as the engineering, because a per-image cost that looks acceptable in testing can be ruinous across a full year of production.
Solution
CodeLeap worked alongside the client over several years to build the platform in full. We began at the source, with iOS and Android capture apps that guide photographers through each vehicle so the images entering the pipeline are consistent by construction. The custom computer vision model then identifies the vehicle and its features, background replacement brings every image to studio standard, and automated quality control inspects the output so that only images meeting the defined standard reach a listing. Approved images feed automatically into interactive 360-degree turntable spins delivered through the platform's web media player.
Underpinning all of it is an infrastructure built to process millions of images a year at a cost per image that keeps the product commercially viable. Quality no longer depends on who happens to be checking, and staff spend their time on the exceptions rather than inspecting each shot.
Tech stack
The tech chosen was our speciality, enabling fast development of performant AI pipelines and a reliable infrastructure.
- Frontend
- React Native (iOS & Android capture apps), React (web media player)
- Backend
- Python, Django, PostgreSQL
- Hosting
- AWS
- AI
- Custom-trained computer vision models
Keep exploring
-
Every listing becomes a video. Automated vehicle video generationA company working in automotive imaging wanted a branded promotional video for every vehicle listing, at a volume no editing team could reach. CodeLeap built the video system end to end, and it has run in production since, producing listing videos with no human editing step.
read more -
SBUSBU is a UK energy broker placing gas and electricity contracts for business customers, in a market where the deal goes to whoever quotes first and follows up hardest. Their chance to win each customer sat in Salesforce as a contract end date with nothing acting on it, so CodeLeap automated everything between that date and the signed contract. Within three months the team was generating five times the qualified leads, without adding headcount.
read more -
WeSnagA major UK construction company ran snagging on spreadsheets, paper checklists and manual follow-ups, which could not scale to dozens of contractors, hundreds of units and thousands of defects. CodeLeap built a platform digitising the whole defect lifecycle, replacing hundreds of hours of manual coordination and leading to further work extending it.
read more
