Supervision: Mohan Vamsi Nallapareddy

Project type: Semester project (bachelor) Semester project (master)


Background: In biomedical research, the detection, counting, and localization of ribosome spots from fluorescent imaging data is crucial for various applications such as particle tracking in live imaging data, and for quantifying mRNAs. For this task, a wide range of techniques from simple image processing filters to supervised deep learning techniques have been applied [1]. Although supervised deep learning techniques have shown promise in this field, the amount of labeled samples has been a limiting step for their progress.

The project: This semester project would explore the application of state-of-the-art unsupervised and self-supervised anomaly detection algorithms specific to the computer vision domain to automate the detection and localization of ribosome spots from microscopy images obtained from various imaging techniques. The student would be working closely with our collaborating lab led by Prof. Jeffrey Chao (FMI Basel) in order to leverage the different kinds of imaging data that is being generated by them.

Required Skills: A basic understanding of neural networks (Multi-Layer Perceptrons, Convolutional Neural Networks, general Machine Learning notions) is required. Prior experience working with deep learning frameworks such as PyTorch or TensorFlow would be highly preferable. Interest in exploring anomaly detection algorithms in the unsupervised / self-supervised learning domain would be a plus.

My background: Vamsi Nallapareddy. I am a Computational Biologist whose primary experience lies in Machine Learning for Structural Biology. I have additionally worked on multiple projects in computer vision for analyzing biomedical data. You can reach me at if you have any questions about the project!

  1. Eichenberger et. al. deepBlink: threshold-independent detection and localization of diffraction-limited spots, Nucleic Acids Research,