Digital Commons Network^™

Effect Of Label Noise On The Machine-Learned Classification Of Earthquake Damage, Jared Frank, Umaa Rebbapragada, James Bialas, Thomas Oommen, Timothy C. Havens

Michigan Tech Publications

Automated classification of earthquake damage in remotely-sensed imagery using machine learning techniques depends on training data, or data examples that are labeled correctly by a human expert as containing damage or not. Mislabeled training data are a major source of classifier error due to the use of imprecise digital labeling tools and crowdsourced volunteers who are not adequately trained on or invested in the task. The spatial nature of remote sensing classification leads to the consistent mislabeling of classes that occur in close proximity to rubble, which is a major byproduct of earthquake damage in urban areas. In this study, …

Bruise Detection In Apples Using 3d Infrared Imaging And Machine Learning Technologies, Zilong Hu

Dissertations, Master's Theses and Master's Reports

Bruise detection plays an important role in fruit grading. A bruise detection system capable of finding and removing damaged products on the production lines will distinctly improve the quality of fruits for sale, and consequently improve the fruit economy. This dissertation presents a novel automatic detection system based on surface information obtained from 3D near-infrared imaging technique for bruised apple identification. The proposed 3D bruise detection system is expected to provide better performance in bruise detection than the existing 2D systems.

We first propose a mesh denoising filter to reduce noise effect while preserving the geometric features of the meshes. …