Real-time visual wrinkle segmentation during paper deep drawing

Abstract Paper deep drawing, inspired by the metal forming technique, is investigated as sustainable alternative to plastic packaging, aiming to achieve airtight sealing and versatile design. However, paper’s inherent tendency to form wrinkles complicates the production process, making it difficult to achieve consistent quality. To effectively control the forming of wrinkles, a runtime-optimized algorithm capable of quickly detecting them for precise material flow manipulation is required. Related fields like facial wrinkle and concrete crack detection provide promising methodologies, particularly using Gabor filters due to their edge-detection capabilities. However, speed optimization has rarely been prioritized in existing studies. Addressing these challenges, this research introduces the Runtime Optimized Wrinkle Detection (ROWDe) algorithm, employing Gabor filters for high accuracy and speed. ROWDe significantly reduces computation time through GPU acceleration, optimized kernel selection, and streamlined pre-processing. Tests on 12-megapixel images yield processing times of under 40 ms, suggesting the method’s suitability for real-time processing. With an average precision of 0.9151 and a recall of 0.8157 on a manually annotated paper deep drawing dataset, the algorithm proves robust against variations in lighting, paper color, and die shape.

Authors

Institutions

Publication Details

Journal
Machine Vision and Applications
Published
2026-11-17
DOI
https://doi.org/10.1007/s00138-026-01953-3
Primary Topic
Industrial Vision Systems and Defect Detection
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Real-time visual wrinkle segmentation during paper deep drawing

Peter Groche, Jan Bechtold, Arjan Kuijper, Arne Rak et al.
3 links
Machine Vision and Applications
Industrial Vision Systems and Defect Detection
article

Real-time visual wrinkle segmentation during paper deep drawing

Peter Groche, Jan Bechtold, Arjan Kuijper, Arne Rak, Volker Knauthe, Tristan Wirth, Cédric Brunk
article en

Abstract

Abstract Paper deep drawing, inspired by the metal forming technique, is investigated as sustainable alternative to plastic packaging, aiming to achieve airtight sealing and versatile design. However, paper’s inherent tendency to form wrinkles complicates the production process, making it difficult to achieve consistent quality. To effectively control the forming of wrinkles, a runtime-optimized algorithm capable of quickly detecting them for precise material flow manipulation is required. Related fields like facial wrinkle and concrete crack detection provide promising methodologies, particularly using Gabor filters due to their edge-detection capabilities. However, speed optimization has rarely been prioritized in existing studies. Addressing these challenges, this research introduces the Runtime Optimized Wrinkle Detection (ROWDe) algorithm, employing Gabor filters for high accuracy and speed. ROWDe significantly reduces computation time through GPU acceleration, optimized kernel selection, and streamlined pre-processing. Tests on 12-megapixel images yield processing times of under 40 ms, suggesting the method’s suitability for real-time processing. With an average precision of 0.9151 and a recall of 0.8157 on a manually annotated paper deep drawing dataset, the algorithm proves robust against variations in lighting, paper color, and die shape.

Machine Vision and ApplicationsVol. 38(1)
Fraunhofer Institute for Computer Graphics Research (DE), Technische Universität Darmstadt (DE)
Deutsche Forschungsgemeinschaft, Technische Universität Darmstadt
Industrial Vision Systems and Defect Detection
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon