• Publications
  • Influence
Adapting Visual Category Models to New Domains
TLDR
We present a method that adapts object models acquired in a particular visual domain to new imaging conditions by learning a transformation that minimizes the effect of domain-induced changes in the feature distribution. Expand
Appearance-based gaze estimation in the wild
TLDR
We present an extensive evaluation of several state-of-the-art image-based gaze estimation algorithms on three current datasets, including our own. Expand
A Multi-World Approach to Question Answering about Real-World Scenes based on Uncertain Input
TLDR
We propose a method for automatically answering questions about images by bringing together recent advances from natural language processing and computer vision. Expand
On the Significance of Real-World Conditions for Material Classification
TLDR
We propose a pure-learning approach which accommodates variations in scale in the training samples, similar to how differing illumination and pose are modelled. Expand
Ask Your Neurons: A Neural-Based Approach to Answering Questions about Images
TLDR
We propose Neural-Image-QA, an approach to question answering with a recurrent neural network. Expand
ML-Leaks: Model and Data Independent Membership Inference Attacks and Defenses on Machine Learning Models
TLDR
We present the most comprehensive study so far on this emerging and developing threat using eight diverse datasets which show the viability of the proposed attacks across domains. Expand
Discovery of activity patterns using topic models
TLDR
In this work we propose a novel method to recognize daily routines as a probabilistic combination of activity patterns. Expand
It’s Written All Over Your Face: Full-Face Appearance-Based Gaze Estimation
TLDR
We propose a full-face CNN architecture for gaze estimation that, in stark contrast to a long-standing tradition, takes the full face image as input and directly regresses to 2D or 3D gaze estimates. Expand
Disentangled Person Image Generation
TLDR
A novel, two-stage reconstruction pipeline that learns a disentangled representation of the aforementioned image factors and generates novel person images at the same time. Expand
A category-level 3-D object dataset: Putting the Kinect to work
TLDR
Recent proliferation of a cheap but quality depth sensor, the Microsoft Kinect, has brought the need for a challenging category-level 3D object detection dataset to the fore. Expand
...
1
2
3
4
5
...