|
 |
News |
|
|
The workshop was cancelled. It will take place later.
|
| General |
|
The workshop "Performance Evaluation of Pattern Recognition in Remote Sensing" of the ISPRS Intercommission Working Group III/VII "Pattern Recognition in Remote Sensing" was planed to be held in Jena, Germany, on 08th September 2009 in conjunction with the DAGM Symposium 2009, but the workshop is now cancelled.
The goal of the workshop is to discuss evaluation methods for image classification in remote sensing.
The great variety of image classification methods available and used requires both a conceptual and an empirical validation
using reference data. As domain knowledge is explicitly or implicitly introduced, the value of such prior information needs
to be evaluated based on commonly agreed performance measures. Benchmarking on standard reference data sets appears to be
the only method to arrive at objective conclusions and at the same time to stimulate research in improving classification
methods. The generating of such benchmarking data sets with real 'ground truth' may be accompanied by reference data set,
where the ground truth is generated by visual or semi-automatic interpretation of images of higher resolution or stereo
imagery. This eases the generation of reference data. Finally, the significance of performance measures needs to be
discussed as they are the basis for proving the difference of classification methods.
It is the goal of the workshop to bring together researchers interested in making the quality of classification methods used in remote sensing transparent and discuss theoretical and practical problems in achieving this goal.
Authors are encouraged to submit papers dealing with but not restricted to the following topics:
- New developments in image classification methods
- Comparison of classification methods
- Value of prior knowledge
- Role of additional data
- Feature selection and extraction
- Spatial, temporal and semantic models
- Incremental learning methods
- Benchmarking and Perfomance measures
- Test scenarios
- Methods for generating reference data sets
|
|
