Computer Vision for Healthcare · Surgical Data Science · Deep Learning

I am a Computer Science master's student at TU Darmstadt and a working student at Fraunhofer IGD. I build computer vision systems for surgical video understanding and medical image analysis, with a focus on reliable evaluation and tools that keep humans in the review loop.

• Surgical video understanding: detecting visual artifacts and segmenting surgical scenes.
• Medical image segmentation: studying U-Net and foundation-model workflows for ophthalmic and histopathology images.
• Review and curation tools: turning frame-level predictions into timelines, editable annotations, and reproducible analysis.

Built a surgical video analysis pipeline using SegSTRONG-C and Cholec80, with a ViT-based classifier for smoke, blood, low brightness, blur, and lens contamination. Implemented frame-level inference, CSV export, timeline visualization, and an interactive GUI for reviewing and correcting predictions.

→ Read the project overview

Fine-tuned SAM2 on CaDIS and studied its integration into SASVi for surgical scene segmentation. The work evaluates spatial segmentation quality and temporal consistency, with qualitative transfer to OphNet.

Nominated for the Freudenberg Award Digital Science 2026, among the top three bachelor's theses at TU Darmstadt.

→ Read the thesis poster

Investigated U-Net segmentation and exploratory cell counting in microscopy images, including data splits, patch-based training, and edge effects. The public notebook implements annotation processing, training, and prediction visualization in TensorFlow/Keras.

→ Explore the notebook

I am interested in exchanging ideas and collaborating on computer vision for healthcare, medical image analysis, and surgical data science.

LinkedIn · Portfolio · Email

_{Visual artwork generated with OpenAI ImageGen and animated specifically for this profile.}