Our Core Technology
Customised Neural Networks
Our customized neural networks (patent pending) specialise in damage detection on real world images with varied angles, lighting conditions etc.
Understanding of Real World Complexities
We have a deep understanding of complexities of detecting damages on physical assets which even confuse human eyes e.g. dent vs reflection, dirt vs scratch.
Data Library
We have trained our neural networks on 7 MN+ damaged physical asset images/videos. As a result, we do not require or store client’s images for training and production.
Video Based Inspections
Unlike other platforms that work only on photos, our platform works on both photos and videos of a car. Videos give a significant boost in accuracy w.r.t our competitors.
Privacy
We mask all personal information before processing images or videos, and do not store any personal data.
Face
License Plate
ID, VIN, etc.
Metadata, eg. GPS Coordinates, Owner
Resilent Infrastructure
Our infrastructure runs on data centers provided by Azure and AWS. We host our solution in the same data center as the client's country
Load Balancing
Scalable DNS
Virtual Private Cloud
Encryption
Access Control
Frequently Asked Questions
Which cloud platforms does Inspektlabs use to host its services?
Inspektlabs currently uses local AWS servers to store data and host its services.
What kind of camera setups are required for automated inspections?
Inspektlabs' AI can be used either via a fixed camera setup or via the cameras on your smartphone
How does Inspektlabs ensure data security and compliance with global regulations?
Inspektlabs is ISO 27001 certified and GDPR compliant to guarantee data secruity.
We also have setup checks to mask PII (personal identifiable information) and can setup scheduled purges for your data upon request.
What kind of models does your damage detection solution use?
Inspektlabs uses a proprietary algorithm that is trained to detect damages on 162 vehicle parts using computer vision.
We use an ensemble of object detection and segmentation model based on different use-cases to detect vehicle damage. You can read more about it on our blog.
