Bioimage computing has known a tremendous development in recent years. State-of-the-art light microscopy (LM) can deliver 2D and 3D image sequences of living cells with unprecedented image quality and ever growing resolution in space and time. The emergence of novel diverse LM modalities has provided biologists with formidable means to explore cell mechanisms, embryogenesis, or neural development, to quote just a few fundamental biological issues. Electron microscopy (EM) supplies information on the cell structure down to the nanometer resolution. Correlating LM and EM at the microscopic level, and both with animal behavior at the macroscopic level, is of paramount importance. In the face of huge data sets, at a size of multiple terabytes per volume or video, and exceedingly difficult problems, state-of-the-art computer vision are required and need to be further developed.
This workshop aims to bring the latest challenges in bioimage computing to the computer vision community. Simultaneously it will showcase the specificities of bioimage computing and its latest achievements. This includes important issues related to image modeling, denoising, super-resolution, semantic and instance segmentation, motion estimation, image registration, tracking, classification, event detection, topics which appertain to the computer vision field.
Topics of interest include but are not limited to:
- LM and EM image restoration and reconstruction
- Reproducible image analyses over terabyte-sized images
- Deep learning for bioimaging
- 3D registration
- Segmentation of subcellular objects, cells, and animals (instances and classes)
- Multimodal image analysis (correlative LM and EM, various LM modalities)
- Analysis of motion: particle tracking and tracking of cells, tissues, and organisms
- Automated behavior recognition
- Dense motion estimation in 2D and 3D LM image sequences
- Diffusion computation
- Statistical analysis of (object) shape
- Image-based phenotyping
- Evaluation and benchmarking methodologies of automated image algorithms/pipelines
- Interactive image analyses (of gigapixel images)
- Creating large collections of labeled training datasets
- other technically interesting and biologically useful pipelines and algorithms…
Papers should adhere to the ICCV 2023 proceedings style requirements and are to be submitted through OpenReview (link will be published here soon). Submission will close on July 21, 2023. The submission instructions of ICCV apply, except that no pre-registration is required and supplemental material is to be submitted together with the main manuscript at the submission deadline. Review will be double-blind.
Please note that your OpenReview paper ID will only be visible after you have uploaded a PDF to OpenReview. In order to have the correct paper ID in your PDF, you will therefore have to upload your paper first with a dummy ID, and then a second time with the correct OpenReview ID.