README.md

In heike/x3prplus: Algorithms for Matching Bullet Lands

x3prplus - Analying Bullet Striations From .x3p Files

The steps of the algorithm and a front-end to the x3prplus R package have been created as a Shiny Web Application. The application can be accessed at https://erichare.shinyapps.io/x3prplus

One can still follow the instructions at the bottom to reproduce the study, but for a comparison of two bullet lands, the web application is a simpler approach.

Introduction

This repository contains the R package x3prplus, along with supporting resources for manipulating bullet images to determine whether two bullets are a match (fired from the same gun barrel) The repository branches are organized as follows:

• master Contains the R package x3prplus
• application Contains a companion Shiny application to assess the probability that two bullets are a match
• application++ Contains a companion Shiny application which walks through each algorithm stage
• papers Contains academic papers involving this package
• presentations Contains presentations involving this package

Installation

The x3prplus package can be installed using devtools:

devtools::install_github("heike/x3prplus")

The Hamby Study

The Hamby study bullets can be assessed using this package, as detailed in Hare, Hofmann, Carriquiry (2016) (See papers/jcgs-2016). To do so, follow these steps:

1. In an R session, install x3pr (Petraco 2015) with install_github("npetraco/x3pr") (See http://www.github.com/npetraco/x3pr for more information)
2. Install x3prplus with devtools::install_github("heike/x3prplus")
3. Install more supporting packages with install.packages(c("dplyr", "ggplot2", "gridExtra", "zoo"))
4. Download the Hamby study images from the NIST Database
   • Browse to https://tsapps.nist.gov/NRBTD
   • Click "Search and Download"
   • Leaving everything default, click Search at the bottom right.
   • Scroll down to the table, and click the "# of Bullets" column twice to sort in descending order. The two Hamby datasets will be the first two options.
   • Check the box next to each, and choose Download Files
5. Unzip the folders of images to a directory of your choosing. For instance, a path such as ~/Desktop/Hamby252_3DX3P1of2 should contain all the images of the form "Br1 Bullet 1-1.x3p", and a path such as ~/Desktop/Hamby252_3DX3P2of2 should contain all the images of the form "Ukn Bullet B-1.x3p"
6. Create a folder csvs in the same location that contains the Hamby folders
7. Execute the code stored in papers/jcgs-2016/code/checkBullets.R Make sure to set the variables knowndatadir and unknowndatadir to the two directories from step 5.
8. Execute the code stored in papers/jcgs-2016/code/complete_hamby.R. Again, make sure to set the variables knowndatadir and unknowndatadir to the two directories from step 5.
9. Execute the code stored in papers/jcgs-2016/code/evaluation_hamby.R.
10. The complete results will be stored in the csvs folder, in the file name bullet-stats.csv

heike/x3prplus documentation built on May 17, 2019, 3:24 p.m.