Automated analysis of leukocyte recruitment for in vivo studies using a spatiotemporal approach and multiple image features

• Autor(a) principal: Silva, Bruno César Gregório da
• Data de Publicação: 2020
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Repositório Institucional da UFSCAR
• Texto Completo: https://repositorio.ufscar.br/handle/ufscar/13137
• topic: Detecção de células; Rastreamento de células; Microscopia intravital; Análise espaço-temporal; Recrutamento leucocitário; Cell detection; Cell tracking; Intravital video microscopy; Spatiotemporal analysis; Leukocyte recruitment; CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::METODOLOGIA E TECNICAS DA COMPUTACAO

Over the last few years, many researchers have directed their efforts and interests toward in vivo studies of the cellular and molecular mechanisms in the microcirculation of many tissues under different inflammatory conditions. These studies’ main goal is to develop more effective therapeutic strategies for the treatment of inflammatory and autoimmune diseases. Leukocyte recruitment analysis is a crucial step to understand the interactions between leukocytes and endothelial cells in the microcirculation of living animals. Performed preferably by the intravital video microscopy (IVM) technique, this procedure usually requires an expert to perform visual analysis, which is prone to the inter- and intra-observer variability, besides being a tedious and time-consuming task. This problem claims, therefore, an automated method to detect and track these cells. To this end, this work aims to study and develop computational techniques for the detection and tracking of leukocytes in IVM images. We proposed an automatic computational pipeline where, after a preprocessing stage, we combined the results of frame-basis detection (2D – spatial processing) with those from three-dimensional analysis (3D=2D+t – spatiotemporal processing) of volumetric images formed by stacking all the video frames. While the 2D processing focuses on leukocytes detection without worrying about their tracking, 2D+t processing was intended to assist in the dynamic analysis of cell movement (tracking). We tested three different detection approaches for the spatial processing, named as MTM-PCA, MTM-DCNN, and DCNN. Our results were obtained by qualitative and quantitative evaluations performed over six different IVM videos, where the detected cells were compared with the manual annotations of an expert. They showed the combination of these both processing stages minimized most of the problems involved in IVM cell detection and tracking, such as cell occlusion and the proper discrimination of cell trajectories.