Selected Publications

We propose to use deep convolutional neural networks to estimate the transient attributes of a scene from a single image. Transient scene attributes describe both the objective conditions, such as the weather, time of day, and the season, and subjective properties of a scene, such as whether or not the scene seems busy. Recently, convolutional neural networks have been used to achieve state-of-the-art results for many vision problems, from object detection to scene classification, but have not previously been used for estimating transient attributes. We compare several methods for adapting an existing network architecture and present state-of-the-art results on two benchmark datasets. Our method is more accurate and significantly faster than all previous methods, enabling real-world applications.

Recent Publications

Recent Posts

A web app with the lighthearted goal of identifying REU sites that satisfy both a students research interests and their musical taste. Suggested sites are defined as those that are within the intersection of states that contain a relevant REU site and states that will host a concert by the relevant artist that summer (e.g. “states that contain micro-biology” ∩ “states that will host Kanye West”). This was written in less than 12 hours as part of the UK ACM’s Fall 2014 Hackathon.

Read more