List All Projects List Graduate Projects List Senior Design Projects List REU Projects Other Undergraduate Projects
Project # 69
A Computer Vision-Powered Toolkit for Analyzing In Vitro Wound Assays
| Team Lead: | Malak Al-Noubani |
| Other Team Members: | Anne Alsup |
| Faculty Advisor(s): | Dr. Michael Cho |
| Department: | BE |
| Project Type: | REU |
| Title: | A Computer Vision-Powered Toolkit for Analyzing In Vitro Wound Assays |
| Abstract: | Wound healing remains a daunting healthcare challenge in the U.S. Developing computational models that mimic or predict tissue repair dynamics will significantly advance wound care. In vitro models are used to induce tissue damage and monitor wound healing over time in response to treatments. Relying on the typical manual analysis of images to quantify wound healing is tedious, laborious, and found to be inaccurate. To enhance image analysis, we developed WoundWare, a pipeline to automate in vitro wound segmentation and calculate wound closure rates by tracking the wound area over time. To further improve WoundWare applicability, we introduce a new, diverse image dataset encompassing various cell types, wound shapes, healing stages, resolutions, and magnifications. The machine learning model's segmentation accuracy is improved with parameter tuning and data cleaning. WoundWare will also be available in ImageJ, a scientific image processing software, in addition to Google Colab for increased accessibility. A new feature introduces healing time estimation based on user-provided images by implementing linear regression in combination with WoundWare's segmentation model. This will allow researchers to monitor wound healing duration without extensive imaging and therefore quickly assess the efficacy of wound treatments.. |
| Poster: | View Poster #69 (PDF) |
| Judges Scoring: | Submit Evaluation (judges only) |