10 - Bimodal Experiments

Question 1

Multimodal imaging

Answer 1

• match adv of each method vz disadv of the other method
• higher range of op°, spatial/temporal coverage, reliability
• lower uncertainty
• robust syst perf

Question 2

Image fusion

Answer 2

• data fusion = combining sensor data (or derived) into a common representational format
• improve info qlty
• strategies
• – multivariate: all data combined using same concept
• – augmented: different fct°s in the fusion process
• – artefact reduction by fusion: 2nd modality used to fill in the blanks
• – combination
• fusion form input or output = data, feature or decision
• workflow = img acq -> registration -> pixel fusion / feature fusion / decision fusion
• registration: manual/guided/automatic

Question 3

Bimodal segmentation

Answer 3

• use bivariate histograms
• classification methods
• Gaussian mixture model
• classification distances (Euclidian, Mahanalubis, Bhattacharia)