![case study](https://cdn.quantiphi.com/2024/03/ed72426f-banner-image-quality-classification.png)
Image Quality Classification
InsuranceBusiness Impacts
87%
Model Accuracy
<5s
Response Time
Customer Key Facts
- Industry : Insurance
- Country : US
Problem Context
The client, a US-based company provides a digital marketplace for Insurance Salvage vehicles. However, due to the bad image quality, customers withdrew their interest in buying vehicles, resulting in a potential loss in revenue. The client, therefore, required a solution that can assess image quality on the go and reject images failing to meet the standards.
Challenges
- Access to the vehicle condition due to the bad quality of images
- Various categories of image noise and different image angles results in many possibilities and complex solution
- Inconsistency in image labeling for training the model
Technologies Used
![Azure DevOps](https://cdn.quantiphi.com/2024/02/Azure-Devops.png)
Azure DevOps
![Azure Machine Learning Service](https://cdn.quantiphi.com/2024/03/3e0582f2-azure-ml.png)
Azure Machine Learning Service
![Azure Repos </br><br>](https://cdn.quantiphi.com/2024/03/8e24350a-azure-repos.png)
Azure Repos
![Azure Blob Storage](https://cdn.quantiphi.com/2024/03/62ee206e-azure-blob-storage.png)
Azure Blob Storage
Developed a Computer Vision Solution Based on the Custom AI/ML Classification Model Technique
Solution
Quantiphi developed a computer vision solution based on the Custom AI/ML Classification Model technique. It identifies image noise such as blurs, glares, shadows, and so on in real-time, prompting field personnel to retake vehicle photos before presenting the car up for bid/sale.
Results
- Developed an automated solution to improvise image quality and remove noise from images
- Increased sales of vehicles as good quality images were submitted to customers
- Ease of image categorization and end to end process in less than 5 seconds