The project is developed in the scope of a digital insurance product. The aim of the AI project is to develop an algorithm to identify a number of furniture elements and from this detection create a referential price of the content identified.

The project is an application that takes an image or a stream. This input is then processed and analyzed by a Deep Learning model. The model used is YOLO v3. Once processing is finished, the output returned by the algorithm is parsed and analyzed. Each element detected is compared with a price table. Finally, after all elements are identified, a summary report is generated with the number of the different elements identified and the total cost of the content detected.

Business Value

Improve the user experience:
– Easy way to encode the content to insure
– Give the possibility to edit the content detected
– The client can receive a personalized offer based on the value of the content to insure
– Automatic report generated summarizing the elements to insure and its value

For the insurer the Digitalization of insurance products open the possibility to acquire new type of clients.

Implementation Plan

– Functional and technical analysis
– Data acquisition
– Development of the Deep Learning model
– Train/ test the model
– Create the application around the model itself
– Setup the REST API system
– Testing
– Deployment of the model on a cloud infrastructure

Commercial representation of YOLO output
Company
Skills
Used technologies
Complexity
Medium
Effort Needed
20 mandays (PoC) and 60 mandays (Prod)
ROI
Diversification in Insurance leads to new markets. The target of a full digital insurance is millennials
Customer
Insurance company
Industry
Insurance
Partner
NaN
Application Type
Image recognition and location
Tech Stack
– Python
– Jupyter
– Pytorch
– OpenCV
– AWS
Learning Path
Github
X
X