Syed Affan Iqbal

Development and Validation of Algorithms for Image-Based Defect Analysis in Stainless Steel Foils for Fuel Cells: - Developed two detection models: - Classical image processing (segmentation & classification) - AI-based neural network approach -Compared both methods regarding accuracy, efficiency & limitations Goal: Enhancing fuel cell quality control through automated defect detection Thanks to Prof. Dr.-Ing. Kristina Kelber & Dr.-Ing. A. Gommlich for their great guidance!

Worked on Apple Blossom Detection with Embedded Computer Vision to optimize fruit yield. - Collected & processed datasets (2D/3D cameras, drones, mobile) - Annotated data (Label Studio, Roboflow) & trained YOLOv8 models - Optimized hyperparameters (Optuna), tracked training (MLflow) - Best model achieved balanced precision/recall for real-world deployment on farming tractors

As a student & researcher, I’ve worked on projects in object detection, augmentation & model optimization. -Enhanced datasets with Albumentations → improved accuracy - Trained & fine-tuned YOLOv8 with Optuna hyperparameter tuning - Developed defect detection algorithms for stainless steel foils (classical CV + AI) - Tech: Python, YOLOv8, Optuna, MLflow, Label Studio, Roboflow