Weekly Update for February 5

Sample Slide

• Do
• not
• trust
• the
• groundhog

Bullet collection

• On Friday we finished collecting bullets from the ten new Sig Sauer hand guns.
• 1,270 bullets and cartridge cases from ten different barrels
• start collecting for ten Smith & Wessons (confiscated by DCI Ankeny)
• scanning

Results from first couple of scans

Bullet lands don’t have very deep marks, but scans look OK

Matching? - not so much

Barrel 1:

On a different note …

Color schemes: CSAFE has it’s own color scheme!

On shoes…

• Study is going well for the most part (looking at you, 3D scanner)
• Planning on extending the period participants will be using the shoes
• Thinking of a follow-up study for the Summer (maybe NIST too)
• Extracting matrices from pressure sensing mat
• Testing ICP algorithm

Groove ID

• RLM method is performing better than rollapply method
  
• Reduces extremity of really bad rollapply examples

Groove ID

• RLM method is performing better than rollapply method
  
• Reduces extremity of really bad rollapply examples

Groove ID

• Also going to try iteratively fitting loess model
  
• Main problem is always ‘how far is too far’ - the cutoff problem

Statistics and Society Book

• Meeting w/Hal via Zoom, then leadership team (outline)
  
• Re-framing Sam’s judge training
• Searching for real-world case studies

REU Results - Hierarchical Model

REU Results - Hierarchical Model

REU Results - Hierarchical Model

REU Results - Hierarchical Model

Data

• Re-install an updated version of Flash ID
• Re-construct and re-export all reference sets (old: REU data, new: CVL database)