ingatlan.com Improves Customer Experience for Uploading Photos Using AWS Machine Learning
ingatlan.com reduces the cost of analyzing 20,000–30,000 photos a day by 60 percent using AWS. ingatlan.com is an online property listings marketplace in Hungary. The company uses Amazon Rekognition and its DetectText feature to ensure that photos comply with its terms of service.
Beyond the savings, AWS is enabling us to enhance security because we no longer send photos for analysis outside of our network to the vendor site.”
Product Manager, Ingatlan.com
A Major Property Listings Marketplace
ingatlan.com, launched in 1997, is the largest property listings website in Hungary with 85 percent of the market share. Homeowners as well as real-estate agencies place ads for properties on ingatlan.com, uploading approximately 20,000–30,000 photos a day.
Text Detection Tool Raises Too Many False Positives
Before any photos appear on the website, ingatlan.com must check that they comply with its terms of service. This means no photos can include text such as the names and contact numbers of the advertisers. ingatlan.com used a third-party text detection tool to identify text in the photos. However, it was expensive to run and ingatlan.com had to send the photos outside its network to a vendor site for analysis, which raised security concerns.
The number of identification errors—false positives—was also high. Tamás Bónis, product manager at ingatlan.com, says, “The situation frustrated customers because once the tool identified possible text, they couldn’t upload the photo until a moderating team had checked the problem.”
Halves False Positives, Saves 20 Hours a Day
ingatlan.com was already using Amazon Web Services (AWS), storing and backing up photos in Amazon Simple Storage Service (Amazon S3), so it looked for a solution among the cloud provider’s services. ingatlan.com then replaced the third-party detection tool with Amazon Rekognition—which automates image analysis with machine learning and its DetectText feature.
“By migrating to Amazon Rekognition, we cut the rate of false positives by half,” says Péter Szabó, product architect at ingatlan.com. With fewer false positives, the moderator team that checks photos picked out by DetectText saves 20 hours a day in needless investigations. Péter says, “It’s given the team time to support new services such as video uploads and virtual property tours.”
Reduces Costs by 60 Percent
Using Amazon Rekognition, ingatlan.com has cut its text detection costs by 60 percent. The savings have enabled the firm to reduce expenditure and create a buffer to help overcome fluctuations in the housing market. Adds Tamás, “Beyond the savings, AWS is enabling us to enhance security because we no longer send photos for analysis outside of our network to the vendor site.”
Hungary-based ingatlan.com is an online property listings website with the largest market share in the country.
Benefits of AWS
- Reduces false positives by 50%
- Saves 20 hours a day in needless investigations
- Cuts text detection costs by 60%
- Focuses more resources on new services
- Enhances data security
AWS Services Used
Amazon Rekognition makes it easy to add image and video analysis to your applications using proven, highly scalable, deep learning technology that requires no machine learning expertise to use. With Amazon Rekognition, you can identify objects, people, text, scenes, and activities in images and videos, as well as detect any inappropriate content.
Amazon Simple Storage Service (Amazon S3) is an object storage service that offers industry-leading scalability, data availability, security, and performance. This means customers of all sizes and industries can use it to store and protect any amount of data for a range of use cases, such as websites, mobile applications, backup and restore, archive, enterprise applications, IoT devices, and big data analytics.
Companies of all sizes across all industries are transforming their businesses every day using AWS. Explore our Machine Learning page and start your own AWS Cloud journey today.