Hi everyone
Instats is excited to offer a 1-day seminar, AI-Powered Image Analysis, livestreaming December 2 and led by Nikolay Oskolkov, Group Leader (PI) at LIOS. Modern microscopy, high-throughput screens, and clinical imaging generate massive datasets that demand automated, quantitative workflows, and this intensive workshop equips you to turn raw images into rigorously validated results. Beginning with the theoretical foundations of convolutional neural networks, you’ll learn how to harness transfer learning, UNet segmentation, Mask R-CNN instance delineation, and object-detection pipelines, while integrating classical OpenCV techniques for hybrid solutions. Nikolay will demonstrate end-to-end coding in Jupyter, using ChatGPT to accelerate development of reproducible Python scripts, data loaders, augmentation routines, training loops, and evaluation dashboards. Special attention is given to Bayesian uncertainty estimation, GRAD-CAM interpretability, and robust benchmarking so that your analyses meet peer-review standards and are ready for theses, grant applications, and high-impact publications. By the end of the day you will be able to design, implement, and critique complete image-analysis experiments: splitting datasets to control batch effects, fine-tuning models on small domain-specific collections, quantifying morphology at single-cell resolution, and documenting every step for transparent sharing.
Sign up today to secure your spot, and feel free to share this opportunity with colleagues and students who might benefit!
Best wishes
Michael Zyphur
Professor and Director
Institute for Statistical and Data Science
Instats is excited to offer a 1-day seminar, AI-Powered Image Analysis, livestreaming December 2 and led by Nikolay Oskolkov, Group Leader (PI) at LIOS. Modern microscopy, high-throughput screens, and clinical imaging generate massive datasets that demand automated, quantitative workflows, and this intensive workshop equips you to turn raw images into rigorously validated results. Beginning with the theoretical foundations of convolutional neural networks, you’ll learn how to harness transfer learning, UNet segmentation, Mask R-CNN instance delineation, and object-detection pipelines, while integrating classical OpenCV techniques for hybrid solutions. Nikolay will demonstrate end-to-end coding in Jupyter, using ChatGPT to accelerate development of reproducible Python scripts, data loaders, augmentation routines, training loops, and evaluation dashboards. Special attention is given to Bayesian uncertainty estimation, GRAD-CAM interpretability, and robust benchmarking so that your analyses meet peer-review standards and are ready for theses, grant applications, and high-impact publications. By the end of the day you will be able to design, implement, and critique complete image-analysis experiments: splitting datasets to control batch effects, fine-tuning models on small domain-specific collections, quantifying morphology at single-cell resolution, and documenting every step for transparent sharing.
Sign up today to secure your spot, and feel free to share this opportunity with colleagues and students who might benefit!
Best wishes
Michael Zyphur
Professor and Director
Institute for Statistical and Data Science
