Image Processing Assistant Notebook - IPAN Combining ImageJ and Python for microscopy image analysis and large-scale data analysis

Abstract

The field of biomedical imaging includes a wide range of techniques used to picture and extract the experimental results as images. The further acquisition of valuable data requires a process of image analysis. ImageJ is one of the most well-known software used in the scientific community to perform image processing analysis. This software has incredible potential because of its numerous features and plugins. However, when the analysis is performed on a large number of data, carrying out the analysis requires the adoption of automatic systems and a certain level of experience. The goal of the following project was to develop an online tutorial and assistant notebook by combining ImageJ functionalities and Python in the Jupyter Notebook platform. For this purpose, the python package PyImageJ has been implemented within python-based functions. Creating reproducible and ready-to-use analytical pipelines potentially reduced the time and effort required to perform the analysis. Moreover, providing this analytical tool help to standardized the processing steps and increased the reproducibility of the image analysis. The series of notebooks presented in this project both provided a learning tool to understand how to functionally operate with the proposed packages and an assistant tool to help the researchers to carry out their analysis. A series of interviews were performed on experts and potential users of this product. This step aimed to understand the user context and the possible applications of the final product of this project. Finally, an evaluation process was performed to assess the accuracy and the efficiency of the developed analytical method. This last evaluating step highlighted the necessity to further investigate the performance of the final product. Nevertheless, despite the complexity and the multiple limitations, the automatization of ImageJ 1.x functionalities throughout Python 3.x was demonstrated to be feasible and potentially improvable.