DIVAS: Outreach to the Natural Sciences through Image Processing

The DIVAS (Digital Imaging and Vision Applications in Science) project addresses workforce challenges in science, technology, engineering, and mathematics by creating a pedagogical and programmatic "on-ramp" that empowers natural science majors to engage in authentic computational problems as members of skilled, professional teams. We are developing and testing institutional practices and curricular innovations that engage and train STEM undergraduate students to use Python programming, and image processing in particular, in their undergraduate research projects. Students are recruited into the DIVAS program in the first semester of their first year. DIVAS scholars and other participating students can experience a variety of interventions including: 1) a one-credit DIVAS seminar exploring several imaging and computing topics; 2) image capture and analysis modules in introductory- or upper-level biology and chemistry courses; 3) a week-long, intensive coding bootcamp that introduces bash, git, Python programming, and the OpenCV image processing library; 4) pair programming exercises to solve genuine morphometric and colorimetric problems; 5) an extended summer research project involving image processing; and 6) weekly code reviews to check on progress and provide guidance. The DIVAS projects measures the impact of these interventions on students' self-reported efficacy in using computation to solve problems, their attitudes towards computation, and their computational thinking skills, using both established and newly developed instruments. Our first year results show that multiple interventions have had significant positive impact on students' self-efficacy and interest in using computing in their future careers, and certain computational thinking skills.