• Publications
  • Influence
GANomaly: Semi-Supervised Anomaly Detection via Adversarial Training
TLDR
Anomaly detection is a classical problem in computer vision, namely the determination of the normal from the abnormal when datasets are highly biased towards one class (normal) due to the insufficient sample size of the other class. Expand
  • 168
  • 34
  • PDF
Real-Time Monocular Depth Estimation Using Synthetic Data with Domain Adaptation via Image Style Transfer
TLDR
We take advantage of style transfer and adversarial training to predict pixel perfect depth from a single real-world color image based on training over a large corpus of synthetic environment data. Expand
  • 113
  • 7
  • PDF
Skip-GANomaly: Skip Connected and Adversarially Trained Encoder-Decoder Anomaly Detection
TLDR
We introduce an unsupervised anomaly detection model, trained only on the normal (non-anomalous) samples in order to learn the normality distribution of the domain, and hence detect abnormality based on deviation from this model. Expand
  • 35
  • 5
  • PDF
DepthComp: Real-time Depth Image Completion Based on Prior Semantic Scene Segmentation
TLDR
We address plausible hole filling in depth images in a computationally lightweight methodology that leverages recent advances in semantic scene segmentation. Expand
  • 23
  • 2
  • PDF
Eliminating the Blind Spot: Adapting 3D Object Detection and Monocular Depth Estimation to 360° Panoramic Imagery
TLDR
We present an approach to adapt contemporary deep network architectures developed on conventional rectilinear imagery to work on equirectangular 360\(^\circ \) panoramic imagery. Expand
  • 20
  • 2
  • PDF
Advances of Soft Computing Methods in Edge Detection
TLDR
Artificial Intelligence (AI) techniques are now commonly used to solve complex and ill-defined problems. Expand
  • 43
  • 2
  • PDF
Style Augmentation: Data Augmentation via Style Randomization
TLDR
We introduce style augmentation, a new form of data augmentation based on random style transfer, for improving the robustness of convolutional neural networks (CNN) over both classification and regression based tasks. Expand
  • 27
  • 1
  • PDF
Generative adversarial framework for depth filling via Wasserstein metric, cosine transform and domain transfer
TLDR
A fully-convolutional generative model is conditioned on the available depth information and full RGB colour information from the scene and trained in an adversarial fashion to complete scene depth. Expand
  • 9
  • 1
  • PDF
Veritatem Dies Aperit - Temporally Consistent Depth Prediction Enabled by a Multi-Task Geometric and Semantic Scene Understanding Approach
TLDR
In this paper, we propose a multi-task learning-based approach capable of jointly performing geometric and semantic scene understanding, namely depth prediction (monocular depth estimation and depth completion). Expand
  • 15
  • PDF
Iris segmentation using an edge detector based on fuzzy sets theory and cellular learning automata.
TLDR
In this paper, an iris segmentation system using edge detection techniques and Hough transforms is presented. Expand
  • 12
  • PDF