Keywords: Cell segmentation, transfer-learning, convlstm, phase-contrast microscopy
TL;DR: Transfer learning and recurrent ConvLSTM units to exploit the temporal information to segment 3D migrating cells in phase contrast microscopy videos
Abstract: This work presents a deep learning-based workflow to segment cancer cells embedded in $3$D collagen matrices and imaged with phase-contrast microscopy under low magnification and strong background noise conditions. Due to the experimental and imaging setup, cell and protrusion appearance change largely from frame to frame. We use transfer learning and recurrent convolutional long-short term memory units to exploit the temporal information and provide temporally stable results. Our results show that the proposed approach is robust to weight initialization and training data sampling.
Registration: I acknowledge that acceptance of this work at MIDL requires at least one of the authors to register and present the work during the conference.
Authorship: I confirm that I am the author of this work and that it has not been submitted to another publication before.
Paper Type: novel methodological ideas without extensive validation
Primary Subject Area: Segmentation
Secondary Subject Area: Transfer Learning and Domain Adaptation
Confidentiality And Author Instructions: I read the call for papers and author instructions. I acknowledge that exceeding the page limit and/or altering the latex template can result in desk rejection.