An advanced image processing approach based on parallel growth and overlap handling to quantify neutrite growth

Methods to assess neurite growth in populations of neuronal cells are required for many applications in biological image analysis. In response to the need for efficient methods to assess neurite growth, we have previously proposed an image processing framework to quantify the number of viable cells and the extent of neurite growth [1]. The approach is based on region growing and uses cost penalties and an upper cost limit to ensure that neurite outgrowth is not overestimated. However, these thresholds need to be defined manually and are set to a fixed value for the entire image. Also, the approach is not able to account for overlapping neurites in dense cell populations. For this reason, we propose two extensions to overcome the aforementioned shortcomings: By growing all regions originating from individual cell nuclei simultaneously, the approach adapts to the underlying microscopy image and doesn’t require manually defined cost limits. An overlap handling is introduced, which is particularly valuable in dense cell populations with overlapping neurites. The results demonstrate that our advanced image processing approach generates results which are even closer to the manual ground truth.