Sprint 3 - Defined Objectives
Project-related
1. Increase Testing
Test more to find better hyper-parameters. Therefore, integrate hyper parameter cross-table.
- Did more testing, came up with alternative test data.
- In Sprint 4: Outsource to friends via deploying product on a web page (and incentivize on German / English Blog article).
2. Data Wrangling
Cutting frames from video to account for the two plants. Consider the integration of a color mask.
- Dropped the frame cutting idea. Did not make sense.
- Consider integrating color mask. J is not satisfied.
3. Statistical Analysis
Correlation, lagged correlation. -> Went for starting with happiness indicator.
- In Sprint 4:
- Research more into OpenCV algorithm functionality.
- Support Happiness statement by including vertical/lateral movement of plant.
4. Result Visualization
Depict results in graphs. [Done with Altair.]
5. Group Analysis
Cut video based on sound patterns into sequences to test for interactions (gender, emotional expression). -> Moved to Sprint 4:
- In Sprint 4:
- Cut videos to length of start ‘sound stimuli’ to end ‘movement of plant completed’ to account for correlations.
- Is this useful?
6. Operational Improvements
Improve slow video processing (make use of 26GB RAM, process on GPU, make progress of video processing visible).
- Integrated timer and starting time.
- In Sprint 4: Create a progress bar.
Non-project related
- Progress on Fluent Python (fundamentals in Python book): in Progress (Working on Chapter 3: Sets and Dicts.)
- Try with different videos from Youtube (not carnivore plants): Done (Paprika, Flat fop).
- Published a blog article on Plantions on www.duerr.se in German and in English.
Links
Project
- Development Files: Private Git with Latest Development
- Public Files: Public Git with Video footage
- Personal Blog article on Emotion Tracker in German
- Personal Blog article on Emotion Tracker in English