In this podcast, Trulia's vice president of engineering discusses the importance of computer vision applications to the website's overall goal of helping buyers find homes.
Computer vision technology is undoubtedly one of the coolest areas of deep learning right now, but commercial applications of the technology have been hard to come by. At the moment, the applications mostly have been relegated to components of social media platforms and other consumer technologies.
But that's starting to change, and one of the companies leading the charge toward enterprise computer vision applications is real-estate listing company Trulia.
In this edition of the Talking Data podcast, we speak with Trulia's vice president of engineering, Deep Varma, to learn more about how the company is working to make computer vision a valuable component of its platform.
In the podcast, Varma talks about why Trulia is so focused on computer vision technology, how the company trains and operationalizes its models, and why its approach to computer vision is somewhat different than other leading tech companies.
Computer vision applications, essentially like other use cases of deep learning, require massive amounts of training data.
Fortunately for Trulia, which has been posting a dozen or so images as part of millions of listings over the past several years, the company is rich in data. This has enabled Trulia's computer vision models to go from relatively simple tasks, like distinguishing between living rooms and kitchens, to more advanced applications, like identifying the type of appliances or countertops in a kitchen.
Listen to this podcast to hear more about how Trulia uses deep learning to build its computer vision technology and how these processes can be used in your enterprise.