See how we are responding to COVID-19 and supporting our employees and customers
Back to all case studies

Computer Vision for Efficient Background Removal

Automated vehicle background removal.

Computer Vision for Efficient Background Removal
Key Details

Automated vehicle background removal.

  • Challenge
    ML application for vehicle photo background removal and avoid human interaction with an image
  • Solution
    A cloud-based solution to process and remove image background. Application of a synthetic background, apply shadows and effects
  • Technologies and tools
    Python, PyTorch, OpenCV

Client

The client is a US-based technology company that produces and distributes a hardware solution that consists of multiple cameras. The product is used by many car dealers across the US. It’s designed to take photos of the vehicles that are later placed on their websites for sale. The business goal of the project was to provide an automated way to remove the image background and deliver processed images ready to be placed on the website.

Computer vision application development requires a lot of expertise and effort thus the client turned to us. The client was looking to hire a team with multi-year expertise in picture analysis software development and background removal services.

Challenge: ML to remove vehicle photo background and avoid human interaction with an image

A vast number of vehicle photos on sale have to be placed at the dealerships’ websites. The photos need to have a more professional presentation without the garage backdrop where the photos were taken.

The client used to work with another vendor prior to giving this project to our team. Their previous vendor failed to deliver something tangible. The InData Labs team approached this work with a dataset collection and proper data labelling. We were challenged to create a solution that does contour analysis for image recognition and background removal. The next step was the creation of an ML model that was trained on the collected dataset. Then a pipeline to automate the entire process was set and as a final step in the project, we re-trained the model and made refinements to ensure excellent accuracy.

Solution: a cloud-based solution to process and remove image background. Application of a synthetic background, apply shadows and effects.

The idea behind the project was about removing the backdrop on photos of vehicles located in a garage to create showroom-like images for car dealers. The client requested the solution to be cloud-hosted. The images are being processed and returned via API. At the same time the data is still being collected and the dataset is enriched. So the accuracy is constantly improving. When working on the project, we were using classic image analysis and detection techniques for robust contour analysis for image recognition. For image contour recognition we used neural networks.

Our team developed our own contour detection algorithm for image processing and an artificial intelligence-based background removal tool.

Computer Vision for Efficient Background Removal

Result: AI-powered background remover

InData Labs, as an image background removal service provider, has aided the client with an automated tool that removes the backdrop of the vehicle photos to speed up the process of the car image previews to be placed at dealership websites. The tool we developed makes vehicle photo capturing very easy, minimizes human interaction and collects images with the same look and feel. Since there is no need for a human-operated backdrop removal work the speed of the car image preparation has increased dramatically.

Tags:
  • Python
  • Computer Vision

Contact InData Labs

Want to start getting value from your data? Fill the form. Click send. Let's talk.

    By clicking Send Message, you agree to our Terms of Use and Privacy Policy.
    This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.