Browsing by Author "Nikolov, Ivan"

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Augmenting Anomaly Detection Datasets with Reactive Synthetic Elements
    (The Eurographics Association, 2023) Nikolov, Ivan; Vangorp, Peter; Hunter, David
    Automatic anomaly detection for surveillance purposes has become an integral part of accident prevention and early warning systems. The lack of sufficient real datasets for training and testing such detectors has pushed a lot of research into synthetic data generation. A hybrid approach by combining real images with synthetic elements has been proven to produce the best training results.We aim to extend this hybrid approach by combining the backgrounds and real people captured in datasets with synthetic elements which dynamically react to real pedestrians and create more coherent video sequences. Our pipeline is the first to directly augment synthetic objects like handbags and suitcases to real pedestrians and provides dynamic occlusion between real and synthetic elements in the images. The pipeline can be easily used to produce a continuous stream of randomized augmented normal and abnormal data for training and testing. As a basis for our augmented images, we use one of the most widely used classical datasets for anomaly detection - the UCSD dataset. We show that the synthetic data produced by our proposed pipeline can be used to make the dataset harder for state-of-the-art models, by introducing more varied and challenging anomalies. We also demonstrate that the additional synthetic normal data can boost the performance of some models. Our solution can be easily extended with additional 3D models, animations, and anomaly scenarios.
  • No Thumbnail Available
    Item
    Creating 3D Asset Variations Through 2D Style Transfer and Generated Texture Maps
    (The Eurographics Association, 2023) Nikolov, Ivan; Singh, Gurprit; Chu, Mengyu (Rachel)
    Generating 3D object variations through style transfer models applied to their textures is an easy way for creating content for games and XR applications. Most workflows focus on either generating albedo textures only without changing the underlying surface and not touching the underlying surface or transforming the 3D object directly, which is computationally and resourceheavy. In this paper, we present an initial exploration of an in-between solution that aims to combine the style transfer of albedo textures with the generation of additional maps, such as normal, displacement, and roughness. The results show that the pipeline can generate variations based on different styles, which would enable the addition of smaller 3D-style surface features to objects without transforming their meshes. The project code and generated textures will be available at https: //github.com/IvanNik17/3D-Assets-From-2D-Style-Transfer.