Track genomics, neuroscience and molecular biology.
Search papers, labs, and topics across Lattice.
This study introduces Doppio, a novel dataset specifically designed for contactless weight estimation of falling particles, focusing on coffee grinding as a practical application. By leveraging deep learning techniques, including both spatial feed-forward networks and recurrent spatio-temporal models, the authors evaluate the predictive accuracy and computational efficiency of various approaches for estimating the cumulative weight of falling coffee grounds. The results indicate that these computer vision models can reliably estimate mass without physical contact, paving the way for innovative measurement solutions in industrial contexts.
Deep learning models can accurately estimate the weight of falling particles, transforming how we approach contactless measurement in industrial applications.
Measuring the mass of powder, including falling particles, is a common task in industrial applications. While scales are effective for static measurements, many applications require contactless sensing, where existing solutions are often costly, application-specific, and technically complex. In this work, we investigate computer vision as a practical alternative for contactless mass estimation. As an accessible real-world case study, we focus on coffee grinding and introduce \emph{Doppio}, a novel video dataset capturing videos of falling ground coffee, paired with precise, per-frame ground-truth weight measurements. To demonstrate contactless measuring, we evaluate deep learning-based approaches ranging from purely spatial feed-forward networks to recurrent spatio-temporal models. These models are analyzed with respect to their predictive accuracy and computational trade-offs. We demonstrate that deep learning-based computer vision models accurately estimate the cumulative weight of falling particles, establishing a solid foundation for future vision-based contactless measurement solutions.
TU Munich