iCub World: Friendly Robots Help Building Good Vision Data-Sets

By Sean Ryan Fanello, Carlo Ciliberto, Matteo Santoro, Lorenzo Natale, Giorgio Metta, Lorenzo Rosasco, Francesca Odone. 2013

31 citations recorded by the source corpus.

In this paper we present and start analyzing the iCub World data-set, an object recognition data-set, we acquired using a Human-Robot Interaction (HRI) scheme and the iCub humanoid robot platform. Our set up allows for rapid acquisition and annotation of data with corresponding ground truth. While more constrained in its scopes -- the iCub world is essentially a robotics research lab -- we demonstrate how the proposed data-set poses challenges to current recognition systems. The iCubWorld data-set is publicly available.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Memory Efficient Experience Replay for Streaming Learning (2018, background): “that contains household objects from ten different object categories [19].”
Coresets via Bilevel Optimization for Continual Learning and Streaming (2020, methodology): “In Figure 1: Data summarization on an imbalanced stream of images created from the iCub World 1.0 dataset [15].”
Lifelong Machine Learning with Deep Streaming Linear Discriminant Analysis (2019, methodology): “s updated, its contents are used to train fully-connected layers in a network. We use ExStream to train the output layer of the network. It achieved state-of-the-art performance on the iCub World 1.0 [35] and CORe50 [36] streaming datasets. • iCaRL – iCaRL [16] is a popular incremental batch learnin”
New Metrics and Experimental Paradigms for Continual Learning (2018, methodology): “0 dataset [2] using ResNet-50 [6] embeddings are shown in Table 1.”
Stream-51: Streaming Classification and Novelty Detection from Videos (2020, background): “0 (iCub-1) [12] and iCub World Transformations (iCub-T) [33].”
How Important Is Weight Symmetry in Backpropagation? (2015, methodology): “iCubWorld dataset[32]: We follow the “categorization” protocol; 10 class categorization task; 600 training and 200 testing samples per category.”