2martens/masterthesis
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Added info about preparation and training

Jim Martens
2019-02-04 13:59:12 +01:00
parent af4841523e
commit db60c2c726
1 changed files with 12 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
Home.md

@ -5,4 +5,15 @@ General Idea of the Thesis:
3. Segment point cloud into separate areas for objects. 3. Segment point cloud into separate areas for objects.
4. a) Feed SSD with 2D data for object. Classify object and calculate classification loss (cross entropy loss). 4. a) Feed SSD with 2D data for object. Classify object and calculate classification loss (cross entropy loss).
b) Feed autoencoder with 2D data for object. Calculate encoding/decoding loss. b) Feed autoencoder with 2D data for object. Calculate encoding/decoding loss.
5. In testing phase the encoding/decoding loss tells us if an object is unknown. 5. In testing phase the encoding/decoding loss tells us if an object is unknown.
Preparation:
1. Clean dataset with dataset_cleaner.py developed in master project
2. specify train/validate/test split
3. reorganize dataset inside these splits according to object class rather than movie/frame (metadata only, no duplicate dataset files!)
4. specify inlier and outlier object classes
Training:
**Independent** training of SSD and Adversarial Autoencoder (AAE). Use ground truth bounding box for AAE.