Descrizione del progetto
Cambiare il modo in cui pensiamo a dati e algoritmi nell’apprendimento automatico per la visione artificiale
Le auto a guida autonoma sembrano essere a portata di mano, in parte grazie al successo degli algoritmi di visione artificiale che sono stati sviluppati per essere gli «occhi» di questi veicoli autonomi. Per muoversi nel mondo, è necessario che i veicoli autonomi comprendano gli oggetti dinamici presenti nella scena, ossia che siano in grado di rilevare, segmentare e seguire vari oggetti in movimento. Oggi la visione artificiale è in grado di soddisfare questa necessità, grazie soprattutto ai progressi compiuti nell’apprendimento profondo. La maggior parte dei metodi si basa su reti neurali convoluzionali addestrate su set di dati su larga scala sotto supervisione, ma questo paradigma è sufficiente per rappresentare la complessità delle nostre strade? Il progetto DynAI, finanziato dal CER, andrà oltre l’apprendimento supervisionato. I ricercatori del progetto concepiranno modelli innovativi di apprendimento automatico in grado di imparare direttamente dai flussi video non etichettati.
Obiettivo
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.
Campo scientifico
- 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
Programma(i)
- HORIZON.1.1 - European Research Council (ERC) Main Programme
Argomento(i)
Meccanismo di finanziamento
ERC - Support for frontier research (ERC)Istituzione ospitante
20124 Milano
Italia