Proptech Labs Problem Statements

Recording:

Proptech Labs Problem Statement Sharing Session

Proptech Labs: Problem Statement 1 – Identifying the Conditions of the Property Areas

Prize: 6-months paid internship for undergraduates or fresh graduates in Sri Lanka

1st Place: $350 USD - Microsoft Funded

Proptech Labs offers a cutting-edge Product for Property Inspections across Australia and New Zealand. During inspections, our diligent agents capture multiple photos of the properties. Currently, when issues arise, such as a Dirty Floor, Damaged Wall, or Scratched Door, agents meticulously inspect the photos and manually tag these conditions.

To streamline this process and enhance efficiency, we are planning to implement an automated tagging system. This system will significantly reduce the time spent on manual tagging by automatically identifying and tagging the conditions of the area. In this exciting challenge, participants will develop a machine learning model tasked with identifying conditions such as Damaged, Dirty, or Scratched within the photos.

Recommended Steps for approaching the problem:
• Build a Machine Learning Model (s) using a proven Technology. You can decide on the dataset to train the model.
• Build a mechanism to easily retrain the model (Probably via an API) with the new images so that the accuracy of the model can be increased
• Extend the functionality to support 360 Images
• Expose an API (S) given uploading an image it should return the locations of the Assets with their category. Example JSON Response below.
{
“Data” : [
{
“Type”: “Damaged”,
“CoordinateX”: “458”,
“CoordinateY” : “1200”
},
{
“Type”: “Scratched”,
“CoordinateX”: “200”,
“CoordinateY” : “190”
}
,{
“Type”: “Dirty”,
“CoordinateX”: “458”,
“CoordinateY” : “1200”
}
]
}
• Deploy the API as either Azure App Service or Function App HTTP Trigger

Hint :
Azure Machine Learning, Azure Vision Studio, Semantic Kernel, Azure Open AI, Azure App Services

Proptech Labs: Problem Statement 2 - Identifying Assets in a Property Area

1st Prize: $350 USD - Microsoft Funded
Prize: 6 months Paid Internship for Undergraduates or Fresh Graduates in Sri Lanka

Proptech Labs offers a cutting-edge solution for property inspections spanning across Australia and New Zealand. Our product streamlines the inspection process by empowering inspection agents to capture multiple photos of properties during their assessments. Presently, agents manually tag various elements such as furniture (beds, chairs), equipment (air conditioners, refrigerators, burners), and doorways in these photos. To enhance efficiency, we aim to automate this tagging process. Our goal is to significantly reduce the time spent on manual tagging by implementing an intelligent system capable of automatically identifying and tagging these elements within the photos. In this exciting challenge, participants will develop a machine learning model tasked with accurately identifying these areas and returning detailed information about the type of assets present along with their locations within the photo.

Recommended Steps for approaching the problem:
• Build a Machine Learning Model (s) using a proven Technology. You can decide on the dataset to train the model.
• Build a mechanism to easily retrain the model (Probably via an API) with the new images so that the accuracy of the model can be increased (Feedback Loop)
• Extend the functionality to support 360 Images
• Expose an API (S) given uploading an image it should return the locations of the Assets with their category. Example JSON Response below.
{
“Data” : [
{
“Type”: “Furniture”,
“CoordinateX”: “458”,
“CoordinateY” : “1200”
},
{
“Type”: “Furniture”,
“CoordinateX”: “200”,
“CoordinateY” : “190”
}
,{
“Type”: “Equipment”,
“CoordinateX”: “458”,
“CoordinateY” : “1200”
},
{
“Type”: “Doorway”,
“CoordinateX”: “345”,
“CoordinateY” : “234”
},
{
“Type”: “Furniture”,
“CoordinateX”: “200”,
“CoordinateY” : “190”
}
,{
“Type”: “Doorway”,
“CoordinateX”: “180”,
“CoordinateY” : “456”
}
]
}
• Deploy the API as either Azure App Service or Function App HTTP Trigger

Hint :
Azure Machine Learning, Azure Vision Studio, Semantic Kernel, Azure Open AI, Azure App Services, Azure Functions

Sponsored by:

Powered by:

Stay connected: