Descripción del proyecto
Cambiar la forma en que pensamos en los datos y los algoritmos en el aprendizaje automático para la visión artificial
Los coches autónomos parecen estar al alcance de nuestra mano, en parte gracias al éxito de los algoritmos de visión artificial, que se han desarrollado para ser los «ojos» de esos vehículos. Para moverse por el mundo, los vehículos autónomos necesitan comprender los objetos dinámicos que los rodean, es decir, detectar, segmentar y seguir múltiples objetos en movimiento. La visión artificial puede ahora abordar con éxito este problema, gracias sobre todo a los avances en el aprendizaje profundo. La mayoría de los métodos se basan en redes neuronales convolucionales entrenadas en conjuntos de datos a gran escala de forma supervisada, pero ¿es este paradigma suficiente para representar la complejidad de nuestras calles? En el proyecto DynAI, financiado por el Consejo Europeo de Investigación, se irá más allá del aprendizaje supervisado. Los investigadores del proyecto diseñarán modelos innovadores de aprendizaje automático que aprendan directamente de secuencias de vídeo sin etiquetar.
Objetivo
Computer vision has become a powerful technology, able to bring applications such as autonomous vehicles and social robots closer to reality. In order for autonomous vehicles to safely navigate a scene, they need to understand the dynamic objects around it. In other words, we need computer vision algorithms to perform dynamic scene understanding (DSU), i.e. detection, segmentation, and tracking of multiple moving objects in a scene. This is an essential feature for higher-level tasks such as action recognition or decision making for autonomous vehicles. Much of the success of computer vision models for DSU has been driven by the rise of deep learning, in particular, convolutional neural networks trained on large-scale datasets in a supervised way. But the closed-world created by our datasets is not an accurate representation of the real world. If our methods only work on annotated object classes, what happens if a new object appears in front of an autonomous vehicle? We propose to rethink the deep learning models we use, the way we obtain data annotations, as well as the generalization of our models to previously unseen object classes. To bring all the power of computer vision algorithms for DSU to the open-world, we will focus on three lines of research: 1-Models. We will design novel machine learning models to address the shortcomings of convolutional neural networks. A hierarchical (from pixels to objects) image-dependent representation will allow us to capture spatio-temporal dependencies at all levels of the hierarchy. 2-Data. To train our models, we will create a new large-scale DSU synthetic dataset, and propose novel methods to mitigate the annotation costs for video data. 3-Open-World. To bring DSU to the open-world, we will design methods that learn directly from unlabeled video streams. Our models will be able to detect, segment, retrieve, and track dynamic objects coming from classes never previously observed during the training of our models.
Ámbito científico
- engineering and technologymechanical engineeringvehicle engineeringautomotive engineeringautonomous vehicles
- natural sciencescomputer and information sciencesartificial intelligencecomputer vision
- engineering and technologyelectrical engineering, electronic engineering, information engineeringelectronic engineeringroboticsautonomous robots
- natural sciencescomputer and information sciencesartificial intelligencemachine learningdeep learning
- natural sciencescomputer and information sciencesartificial intelligencecomputational intelligence
Programa(s)
- HORIZON.1.1 - European Research Council (ERC) Main Programme
Régimen de financiación
ERC - Support for frontier research (ERC)Institución de acogida
20124 Milano
Italia