English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Open

Presents the pdf analysis as premium feature

PDF Analysis

Insights

Presents the summarisation as premium feature

Summarisation

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

We report on an extensive study of the benefits and limitations of currentdeep learning approaches to object recognition in robot vision scenarios,introducing a novel dataset used for our investigation. To avoid the biases incurrently available datasets, we consider a natural human-robot interactionsetting to design a data-acquisition protocol for visual object recognition onthe iCub humanoid robot. Analyzing the performance of off-the-shelf modelstrained off-line on large-scale image retrieval datasets, we show the necessityfor knowledge transfer. We evaluate different ways in which this last step canbe done, and identify the major bottlenecks affecting robotic scenarios. Bystudying both object categorization and identification problems, we highlightkey differences between object recognition in robotics applications and inimage retrieval tasks, for which the considered deep learning approaches havebeen originally designed. In a nutshell, our results confirm the remarkableimprovements yield by deep learning in this setting, while pointing to specificopen challenges that need be addressed for seamless deployment in robotics.

Are we done with object recognition? The iCub robot's perspective