Sep 13 - 16 2022 | Rome, Italy

Hyperspectral Sensing meets Machine Learning and Pattern Analysis

Sept. 15, 2022
Rome, Italy

General Chair
Sina Keller

General Chair
Martin Weinmann

Welcome: 

HyperMLPA is an interdisciplinary workshop which aims at bringing together people of different communities and disciplines involved in hyperspectral sensing, machine learning, and pattern analysis. People are invited to contribute in sensor development and calibration, to present and publish new datasets, to present innovative methodological advancements, and to share the latest results and findings of application-oriented work.

Information

The HyperMLPA workshop will run for the third time in 2022, as part of the SpectraExpo in parallel with Whispers and SwIImS. HyperMLPA participants will have full access to all Whispers/SwIImS sessions and vice versa. Proceedings will be published on IEEE Xplore following peer review.

Sensor Design and Calibration

• Sensor and system calibration and validation
• Error modeling
• Data evaluation

Datasets

• New hyperspectral benchmark datasets
• Ground-based, airborne and spaceborne data acquisition
• Datasets acquired in controlled environments (e.g. in a laboratory) vs. datasets acquired “in the wild”

Methodology

• Feature extraction
• Feature selection / dimensionality reduction
• Spectral unmixing
• Data fusion
• Change detection
• Classification, segmentation and regression

Applications

• Land cover and land use classification
• Material classification
• Biodiversity assessment
• Estimation of geo- and bio-physical parameters
• Fusion of hyperspectral, geo- and bio-physical parameters
• Analysis of spatial and temporal patterns

Location

The conference will take place within the amazing historical complex of San Pietro in Vincoli in the facilities of the Engineering School, University “Sapienza” of Rome.

San Pietro in Vincoli, Rome
via eudossiana 18, Roma, Italy

In order to be inclusive, the event will be held in hybrid format. Those unable to attend physically are welcome to participate virtually via our online platform.