In this paper we address the problem of recognising the Broad-leaved dock (Rumex obtusifolius L.) in grasslands from high-resolution 2D images. We discuss and present the determining factors for developing and implementing weed visual recognition algorithms using deep learning. This analysis, leads to the formulation of the proposed algorithm. Our implementation exploits Transfer Learning techniques for deep learning-based feature extraction, in combination with a classifier for weed recognition. A prototype robotic platform has been used to make available an image dataset from a dairy farm containing broad-leaved docks. The evaluation of the proposed algorithm on this dataset shows that it outperforms competing weed/plant recognition methods in recognition accuracy, while producing low false-positive rates under real-world operation conditions.

No abstract

No abstract

No abstract

No abstract

No abstract

This paper presents a natural user interface system based on multimodal human computer interaction, which operates as an intermediate module between the user and the operating system. The aim of this work is to demonstrate a multimodal system which gives users the ability to interact with desktop applications using face, objects, voice and gestures. These human behaviors constitute the input qualifiers to the system. Microsoft Kinect multi-sensor was utilized as input device in order to succeed the natural user interaction, mainly due to the multimodal capabilities offered by this device. We demonstrate scenarios which contain all the functions and capabilities of our system from the perspective of natural user interaction.

No abstract