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/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Object detection is a fundamental task for robots to operate in unstructuredenvironments. Today, there are several deep learning algorithms that solve thistask with remarkable performance. Unfortunately, training such systems requiresseveral hours of GPU time. For robots, to successfully adapt to changes in theenvironment or learning new objects, it is also important that object detectorscan be re-trained in a short amount of time. A recent method [1] proposes anarchitecture that leverages on the powerful representation of deep learningdescriptors, while permitting fast adaptation time. Leveraging on the naturaldecomposition of the task in (i) regions candidate generation, (ii) featureextraction and (iii) regions classification, this method performs fastadaptation of the detector, by only re-training the classification layer. Thisshortens training time while maintaining state-of-the-art performance. In thispaper, we firstly demonstrate that a further boost in accuracy can be obtainedby adapting, in addition, the regions candidate generation on the task at hand.Secondly, we extend the object detection system presented in [1] with theproposed fast learning approach, showing experimental evidence on theimprovement provided in terms of speed and accuracy on two different roboticsdatasets. The code to reproduce the experiments is publicly available onGitHub.

Fast Region Proposal Learning for Object Detection for Robotics