Pattern Recognition

This book constitutes the refereed proceedings of the 47th DAGM German Conference on Pattern Recognition, DAGM-GCPR 2025, held in Freiburg, Germany, during September 23-26, 2025.

The 40 full papers presented in this volume were carefully reviewed and selected from 85 submissions.

They are grouped into the following topics: Computer Vision Systems and Applications; Video Analysis and Synthesis; Machine Learning Methods; Applications of Foundation Models; Safety and Robustness; 3D Perception and Reconstruction; Photogrammetry and Remote Sensing.

Klappentext

.- Computer Vision Systems and Applications. .- Box it and Track it: A Weakly Supervised Framework for Cell Tracking. .- A Cascaded Dilated Convolution Approach for Mpox Lesion Classification. .- HistDiST: Histopathological Diffusion-based Stain Transfer. .- Deep Learning-Assisted Dynamic Mode Decomposition for Non-resonant Background Removal in CARS Spectroscopy. .- γ-Quant: Towards Learnable Quantization for Low-bit Pattern Recognition. .- EVCS: A Benchmark for Fine-Grained Electric Vehicle Charging Station Detection. .- Video Analysis and Synthesis. .- SegSLR: Promptable Video Segmentation for Isolated Sign Language Recognition. .- Video Object Segmentation-aware Audio Generation. .- MCUCoder: Adaptive Bitrate Learned Video Compression for IoT Devices. .- VisualChef: Generating Visual Aids in Cooking via Mask Inpainting. .- StorySync: Training-Free Subject Consistency via Region Harmonization. .- Structured Universal Adversarial Attacks on Object Detection for Video Sequences. .- Road Obstacle Video Segmentation. .- Machine Learning Methods. .- Don’t Miss Out on Novelty: Importance of Novel Features for Deep Anomaly Detection. .- LADB: Latent Aligned Diffusion Bridges for Semi-Supervised Domain Translation. .- On the Dangers of Bootstrapping Generation for Continual Learning and Beyond. .- Combined Image Data Augmentations diminish the benefits of Adaptive Label Smoothing. .- Efficient Masked Attention Transformer for Few-Shot Classification and Segmentation. .- Applications of Foundation Models. .- Using Knowledge Graphs to harvest datasets for efficient CLIP model training. .- Unlocking In-Context Learning for Natural Datasets Beyond Language Modelling. .- Investigating Structural Pruning and Recovery Techniques for Compressing Multimodal Large Language Models: An Empirical Study. .- Assessing Foundation Models for Mold Colony Detection with Limited Training Data. .- Common Data Properties Limit Object-Attribute Binding in CLIP. .- subCellSAM: Zero-Shot (Sub-)Cellular Segmentation for Hit Validation in Drug Discovery. .- Safety and Robustness. .- synth-dacl: Does Synthetic Defect Data Enhance Segmentation Accuracy and Robustness for Real-World Bridge Inspections?. .- FedPCE: Federated Personalized Client Embeddings for Post-training Knowledge Distillation. .- Object Risk Estimation for Autonomous Driving Safety. .- Rethinking Semi-supervised Segmentation Beyond Accuracy: Robustness and Reliability. .- Detection of Synthetic Face Images: Accuracy, Robustness, Generalization. .- 3D Perception and Reconstruction. .- MT-Occ: Single-View 3D Occupancy Prediction via Multi-Task Distillation. .- Hierarchical Insights: Exploiting Structural Similarities for Reliable 3D Semantic Segmentation. .- CoProU-VO: Combining Projected Uncertainty for End-to-End Unsupervised Monocular Visual Odometry. .- Combining Absolute and Semi-Generalized Relative Poses for Visual Localization. .- Graph Roof Reconstruction with Synthetic Data from Misaligned Labels. .- sshELF: Single-Shot Hierarchical Extrapolation of Latent Features for 3D Reconstruction from Sparse-Views. .- Photogrammetry and Remote Sensing. .- NaT-ReX: Naturalness Assessment with Transformer-Based Reliable Explainability. .- Semantic Segmentation of Structural Damage: A Comparative Study of YOLO11 and Encoder-Decoder Networks. .- Can Multitask Learning Enhance Model Explainability?. .- Out-of-Distribution Detection in LiDAR Semantic Segmentation Using Epistemic Uncertainty from Hierarchical GMMs. .- RadarSeq: A Temporal Vision Framework for User Churn Prediction via Radar Sequence Chart.

Inhalt

.- Computer Vision Systems and Applications.

.- Box it and Track it: A Weakly Supervised Framework for Cell Tracking.

.- A Cascaded Dilated Convolution Approach for Mpox Lesion Classification.

.- HistDiST: Histopathological Diffusion-based Stain Transfer.

.- Deep Learning-Assisted Dynamic Mode Decomposition for Non-resonant Background Removal in CARS Spectroscopy.

.- -Quant: Towards Learnable Quantization for Low-bit Pattern
Recognition.

.- EVCS: A Benchmark for Fine-Grained Electric Vehicle Charging
Station Detection.

.- Video Analysis and Synthesis.

.- SegSLR: Promptable Video Segmentation for Isolated Sign Language
Recognition.

.- Video Object Segmentation-aware Audio Generation.

.- MCUCoder: Adaptive Bitrate Learned Video Compression for IoT Devices.

.- VisualChef: Generating Visual Aids in Cooking via Mask Inpainting.

.- StorySync: Training-Free Subject Consistency via Region Harmonization.

.- Structured Universal Adversarial Attacks on Object Detection for
Video Sequences.

.- Road Obstacle Video Segmentation.

.- Machine Learning Methods.

.- Don't Miss Out on Novelty: Importance of Novel Features for Deep
Anomaly Detection.

.- LADB: Latent Aligned Diffusion Bridges for Semi-Supervised Domain
Translation.

.- On the Dangers of Bootstrapping Generation for Continual Learning
and Beyond.

.- Combined Image Data Augmentations diminish the benefits of
Adaptive Label Smoothing.

.- Efficient Masked Attention Transformer for Few-Shot Classification
and Segmentation.

.- Applications of Foundation Models.

.- Using Knowledge Graphs to harvest datasets for efficient CLIP model
training.

.- Unlocking In-Context Learning for Natural Datasets Beyond Language
Modelling.

.- Investigating Structural Pruning and Recovery Techniques for
Compressing Multimodal Large Language Models: An Empirical Study.

.- Assessing Foundation Models for Mold Colony Detection with Limited
Training Data.

.- Common Data Properties Limit Object-Attribute Binding in CLIP.

.- subCellSAM: Zero-Shot (Sub-)Cellular Segmentation for Hit Validation
in Drug Discovery.

.- Safety and Robustness.

.- synth-dacl: Does Synthetic Defect Data Enhance Segmentation
Accuracy and Robustness for Real-World Bridge Inspections?.

.- FedPCE: Federated Personalized Client Embeddings for Post-training
Knowledge Distillation.

.- Object Risk Estimation for Autonomous Driving Safety.

.- Rethinking Semi-supervised Segmentation Beyond Accuracy:
Robustness and Reliability.

.- Detection of Synthetic Face Images: Accuracy, Robustness, Generalization.

.- 3D Perception and Reconstruction.

.- MT-Occ: Single-View 3D Occupancy Prediction via Multi-Task Distillation.

.- Hierarchical Insights: Exploiting Structural Similarities for Reliable 3D
Semantic Segmentation.

.- CoProU-VO: Combining Projected Uncertainty for End-to-End
Unsupervised Monocular Visual Odometry.

.- Combining Absolute and Semi-Generalized Relative Poses for Visual
Localization.

.- Graph Roof Reconstruction with Synthetic Data from Misaligned Labels.

.- sshELF: Single-Shot Hierarchical Extrapolation of Latent Features for
3D Reconstruction from Sparse-Views.

.- Photogrammetry and Remote Sensing.

.- NaT-ReX: Naturalness Assessment with Transformer-Based Reliable
Explainability.

.- Semantic Segmentation of Structural Damage: A Comparative Study
of YOLO11 and Encoder-Decoder Networks.

.- Can Multitask Learning Enhance Model Explainability?.

.- Out-of-Distribution Detection in LiDAR Semantic Segmentation Using
Epistemic Uncertainty from Hierarchical GMMs.

.- Ra…