1 / 18

Enabling Users to Guide the Design of Robust Model Fitting Algorithms

Learn how to design objective functions for accurate and efficient model fitting algorithms using our 5-step machine learning approach. Improve accuracy and reduce manual steps in model-based image interpretation.

clementv
Download Presentation

Enabling Users to Guide the Design of Robust Model Fitting Algorithms

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Matthias Wimmer, Freek Stulp and Bernd Radig matthias.wimmer@cs.tum.edu Technische Universität München Enabling Users to Guide the Design of Robust Model Fitting Algorithms

  2. Outline • Model-based image interpretation • Model fitting, objective function • Designing objective functions • Our 5-step approach • Learning objective functions • Partly automated • Evaluation • Accuracy • Runtime

  3. Model-based Image Interpretation • The model The model contains a parameter vector that represents the model’s configuration. video Dvideo U

  4. Model Fitting • Objective functionCalculates a value that indicates how accurately a parameterized model matches an image. • Fitting algorithmSearches for the modelparameters that describe the image best,i.e it minimizes the objective function.

  5. Introducing Objective Functions

  6. Ideal Objective Functions P1: Correctness property:The global minimum corresponds to the best model fit. P2: Uni-modality property:The objective function has no local extrema. ¬ P1 P1 ¬P2 P2

  7. Design Approach Shortcomings: • Many manual steps • Requires domain knowledge • Time-consuming (because of loop) • Low accuracy

  8. Our Approach bases on Machine Learning • Ideal objective function necessary • Distance between current and correct location of contour point • Provides training data • Machine Learning yields calculation rules • Guided by human experience (widely automated) x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x

  9. Step 1: Manually Annotate Images

  10. Step 2: Generate Further Annotations ……...............………………………….. function value = 0.3 function value = 0 function value = 0.2

  11. Step 3: Specify Image Features Styles (6): Sizes (3): Locations (5x5): Number of features: 6 styles · 3 sizes · 25 locations = 450

  12. Step 4: Generate Training Data • Mapping of feature values to the expected function value.

  13. Step 5: Apply Machine Learning Machine learning technique: Model Trees • Select the most relevant features • High runtime performance

  14. Benefits • Locally customized calculation rules • Automatic selection of relevant features • Generalization from many images

  15. Evaluation 1: Fitting Accuracy on BioID

  16. Evaluation 2: Runtime Characteristics statistics-based objective function f m learned objective function f l A: 45.1 ms B: 1360 ms C: 8.12 ms D: 9.75 ms • f m considers all features provided. • f l selects the most appropriate features. • Note: C and D are as accurate as B.

  17. Ongoing Research and Outlook • Integration of further image features • Compute the image features on the fly • Learning objective functions for 3D models • Application to different scenario • Medical scenario • Robot scenario: • Model of indoor environment • Self localization

  18. Thank you! ありがとう Online-Demonstration: http://www9.cs.tum.edu/people/wimmerm

More Related