Scientific Section
Accomplishments:
Problems:
Engineering Section
Accomplishments:
Did some more design on language analysis tool (TAUNT)
Installed/configured phpMyAdmin on playground to allow web administration of study DB (this should make it much more user-friendly to correct the missing scenario/role numbers issue, make sure the description data is formatted consistently, and create a separate table to store just the description [...]
Scientific Section
Accomplishments:
Problems:
Engineering Section
Accomplishments:
TAUNT will now load the study 1 data, allow data subset selection based on scenario/role, and create word frequency plots
Created word frequency plots based on study 1 data for each scenario X {text, video} pair.
Created a web page in the GIG2008 repository to display the language analysis results (including word clouds [...]
Scientific Section
Accomplishments
Problems:
Need to finish work on TAUNT and start performing detailed analysis of study data
Engineering Section
Accomplishments:
Worked through several tasks with NLTK, becoming more familiar with the API (figured out how to extract the data I want into NLTK, create word frequency plots, find interesting bigrams and trigrams)
Began working on TAUNT (Textual Analysis Using [...]
To Do:
Create single-word frequency histograms from Study 1 Data
Chunk Study 1 Data into 2-word pairs and create histograms to see what combinations come up a lot (such as “authorized person” and “unauthorized person”)
Use POS tagging and create word frequency histograms based only on verbs
Generate WordNet synsets for common verbs
Finally, look for clusters in [...]
Scientific Section
Accomplishments:
Started lit search for papers dealing with content generation from unstructured textual descriptions
Problems:
Engineering Section
Accomplishments:
Extracted textual scenario descriptions from Study 1 Data and created word clouds from them
Problems:
Need to become much more proficient with Python/NLTK
Need to finish developing tool chain for extracting information from the study data format
Other
Completed CITI Program IRB training (Completion Report)
Still waiting [...]
Title: Generating a 3D simulation of a car accident from a written description in natural language: The Carsim system (Proceedings of the workshop on Temporal and spatial information processing 2001)
Authors: Dupuy, S. and Egges, A. and Legendre, V. and Nugues, P
Abstract:
This paper describes a prototype system to visualize and animate 3D scenes from car [...]
Title: WordsEye: an automatic text-to-scene conversion system (in SIGGRAPH 2001)
Authors: Bob Coyne & Richard Sproat
Abstract:
Natural language is an easy and effective medium for describing visual ideas and mental images. Thus, we foresee the emergence of language-based 3D scene generation systems to let ordinary users quickly create 3D scenes without having to learn special [...]
Need to download and play around with the GPL’d Python NLP package Natural Language Toolkit (http://www.nltk.org)