Motivation:e scratch assay is a standard experimental protocol used to characterize cell migration. It can be used to identify genes that regulate migration and evaluate the e cacy of potential drugs that inhibit cancer invasion. In these experiments, a scratch is made on a cell monolayer and recolonisation of the scratched region is imaged to quantify cell migration rates. A drawback of this methodology is the lack of its reproducibility resulting in irregular cell-free areas with crooked leading edges. Existing quanti cation methods deal poorly with such resulting irregularities present in the data. Results: We introduce a new quanti cation method that can analyse low quality experimental data. By considering in-silico and in-vitro data, we show that the method provides a more accurate statistical classi cation of the migration rates than two established quanti cation methods. e application of this method will enable the quanti cation of migration rates of scratch assay data previously unsuitable for analysis. Availability and Implementation: e source code and the implementation of the algorithm as a GUI along with an example dataset and user instructions, are available in https://bitbucket.org/ anavictoria-ponce/local migration quantification scratch assays/src/ master/. e datasets are available in https://ganymed.math.uni-heidelberg.de/ ∼victoria/publications.shtml.