Source-linked AI summary
StitchOver: Technical Embroidery on Seamed Fabrics
Zekun Chang, Tianhong Catherine Yu, Yixuan Gao, Thijs Roumen
TL;DR
Seams and machine variability make technical embroidery on real-world textile goods prone to failure, despite its ability to add functionality to existing textiles. StitchOver uses JumpStitches to route conductive threads around seams, and evaluation reports eliminated defects compared with frequent failures without them, including under tension variation.
Problem
Seams and heterogeneous textile construction create tension and friction changes that make reliable technical embroidery difficult, while avoiding seams limits the available area for functional integration.
Source: §1 · p. 2Method
StitchOver automatically adjusts needle penetration points around seams and adds supporting stitches through its JumpStitch mechanism.
Source: §1 · p. 2Results
JumpStitches eliminated defects compared with 7/10 defective samples without them and improved performance under tension variation while maintaining conductivity.
Source: §1 · p. 2Takeaways & Limitations
The evaluation demonstrates a substantial decrease, and mostly elimination, of fabrication defects caused by seams using JumpStitches.
Source: §10 · p. 11The evaluation focused on thick leather and did not systematically characterize embroidery parameters across diverse fabric combinations.
Source: §8.3 · p. 10
Abstract
from arXiv · showhide
Smart textiles embed interactivity into everyday garments, supporting use cases like always-available sensing for medical applications or sports. Machine embroidery allows integrating functionalities into existing textiles. However, embroidering onto real-world textile goods remains challenging. Textile goods are rarely made of a single homogeneous substrate of fabric, and embroidery with functional materials such as conductive threads requires machines to be more tightly calibrated than for decorative embroidery. In particular, seams, which bring together different substrates, along with machine variability, cause shifts in tension and friction between the functional thread and the textile substrate that frequently lead to defects (70% of samples in our evaluation). We present a technique to reliably embroider on seamed fabric even when using functional threads. Our software tool automatically digitizes user-defined stitch patterns by introducing what we call "JumpStitches" to bypass seam interference. We evaluated our approach under varying machine states (under-tensioned, well-calibrated, and over-tensioned), and across multiple seam and pattern configurations. Our results show that the JumpStitch mechanism eliminates defects, while maintaining conductivity compared to 70% defects without JumpStitches, and even in poorly calibrated machine states continues to work well.
1 Introduction
Seams and machine variability make technical embroidery on real-world textile goods prone to defects, while avoiding seams limits where functionality can be integrated. StitchOver uses JumpStitches to adapt stitch placement around seams, and evaluation reports eliminated defects compared with frequent failures without them.
1 Introduction
- Textile goods commonly join multiple materials and panels at seams, so embroidery methods for existing goods must accommodate structural variation.The approach aims to work across seams assuming difficult-to-hoop regions are handled with established embroidery practices.Source: §1 · p. 2
- Seams create local tension and friction changes that can cause fraying, bird nesting, needle breakage, or machine damage, while avoiding seams limits embroidery area.Functional threads are already more sensitive to tension variation and surface friction than conventional threads.Source: §1 · p. 2
- StitchOver inserts JumpStitches that keep conductive thread from penetrating seams and add supporting stitches to secure it.The tool analyzes the relationship between conductive routing and seams, then adjusts needle penetration points and generates supporting stitches.Source: §1 · p. 2
- JumpStitches eliminated defects, compared with 7/10 defective samples without them, and improved performance under tension variation while maintaining conductivity.The evaluation covered variations in machine tension.Source: §1 · p. 2
2 Background
Technical embroidery integrates sensing or actuation into existing textiles, but stitching failures can compromise the intended function. Continuous thread paths matter for many functional materials, and seams are common structural features of real textile goods.
2 Background
- Technical embroidery integrates sensing or actuation into existing textiles using a computerized embroidery machine.Failures such as thread breakage or stitching errors can undermine intended functionality.Source: §2 · p. 2
- Many technical embroidery applications depend on uninterrupted thread paths, because breakage directly causes functional failure.Examples include conductive circuits, shape-transformation threads, liquid-metal tubes, and reinforcement structures.Source: §2 · p. 2
- Seams are common in textile goods and often mark structurally or functionally meaningful regions that experience motion or strain.Examples include shoulder, underarm, and side seams.Source: §2 · p. 2
3 Related Work
Related work spans textile integration, material-aware fabrication, robustness to manufacturing variation, and digitization. StitchOver builds on these areas by adapting embroidery paths to seams and other textile variation.
3 Related Work
- Prior work on textile fabrication and robust methods motivates this paper’s focus on reliable technical embroidery for assembled soft goods.The related work is organized around textile functionality, material variability, robust fabrication, and digitization.Source: §3 · p. 2
3.1 Integrating functionalities into textiles
- Machine embroidery can add functional threads to existing products with precise stitch placement, extending prior textile work on sensing, interaction, haptics, and speakers.E-Sewing also describes stitch strategies for integrating components, terminating connections, and insulating conductive thread.Source: §3.1 · pp. 2–3
3.2 Material variability
- Prior material-variability work detects or characterizes differences and guides fabrication choices; StitchOver applies material-aware fabrication to embroidery by generating parameters and design adjustments.Examples include computer-vision inspection and tools that adapt designs or fabrication settings to material constraints.Source: §3.2 · p. 3
3.3 Robust design
- Robust-design approaches embed tolerance-aware mechanisms so systems remain functional despite variation across machines and fabrication conditions.This work explores applying those principles to machine embroidery.Source: §3.3 · p. 3
3.4 Digitization in fabrication
- Embroidery digitization converts designs into stitch paths and machine parameters, while existing software adjustments generally assume homogeneous materials across a design.This work adapts stitch paths to structural variation while retaining the same design and functional threads.Source: §3.4 · p. 3
4 Mechanism design of JumpStitches
This section explains how seams destabilize technical embroidery and presents stitch modifications that bypass risky regions while adding support where needed.
4.2 Our proposed solution: JumpStitches
- JumpStitches modify stitch patterns to cross seams without defects or functional failure by skipping risky-region penetrations and securing the larger jump with alignment stitches.The jump avoids seam interference; parallel alignment stitches keep the loose conductive thread in place.Source: §4 · p. 3, §4.2 · pp. 4–5
4.1 The problems of embroidering across seams
- A seam lifts the presser foot, increasing friction and destabilizing the fabric; these altered stitching conditions can produce thread breakage, fraying, bird nesting, and pattern distortion.Figure 2 contrasts a presser foot pressed down on flat fabric with one lifted by a seam; Figure 3 shows typical resulting defects.Source: §4.1 · pp. 3–4, Figure 2 · p. 4, Figure 3 · p. 4
- The risky regions are S, the overlapping seam; L, its lower side; H, its higher side; and E, the boundary between S and H, each with distinct sources of risk.L risks incomplete presser-foot contact, H has a hollow beneath it, and E has an abrupt thickness change.Source: §4.1 · p. 4, Figure 4 · p. 4
4.2 Our proposed solution: JumpStitches
- JumpStitches can remove shape-defining needle points and break circuitry, while nonlinear deformation at multiple seam intersections can make them fail.The authors propose ShortStitches and UnderlayStitches to address these limitations.Source: §4.2 · p. 5
- When shape-defining points fall in risky regions, localized ShortStitches use a 0.5 mm stitch length to make presser-foot height transitions more gradual.The smaller XY steps reduce incremental Z displacement as the presser foot encounters thickness changes.Source: §4.2 · p. 5, Figure 6 · p. 5
- At multi-way seam intersections, UnderlayStitches flatten and stabilize overlapping fabric layers to provide consistent support as the embroidery path crosses the junction.Figure 6 depicts the fabric sides and schematic for the intersection stitches.Source: §4.2 · p. 5, Figure 6 · p. 5
5 StitchOver
StitchOver converts a user-aligned stitch pattern and seam layout into a modified plan that avoids risky seam crossings. It also supports larger patterns through multi-hooping and adapts stitch spacing to different fabric materials.
5 StitchOver
- Users align a stitch pattern with a photograph of hooped fabric and define seams; StitchOver then adjusts needle points and exports the modified plan.Seams can be specified by start and end points and direction, including curved seams.Source: §5 · p. 5, Figure 7 · p. 5
- StitchOver warns when path endpoints fall in risk regions and suggests geometric adjustments when consecutive shape-defining points share a risk region.For stitch-level optimization, it inserts ShortStitches around shape-defining points and applies JumpStitches to non-geometry-critical seam-crossing segments.Source: §5 · pp. 5–6
- For patterns larger than a hoop, users can split the design across customized hoops, with connector JointStitches reconnecting interrupted paths.Default zigzag lock stitches can maintain continuity for conductive threads with exterior coating.Source: §5 · p. 6, Figure 8 · p. 6
- StitchOver adjusts needle-point spacing and JumpStitch length according to the material properties assigned to fabric regions.The JumpStitch lengths are proportional to the thickness of the two materials joined by a seam.Source: §5 · p. 6
- The minimum supported jump length is physically constrained by presser-foot dimensions, seam width, and the angle between the thread path and presser foot.The minimum length is set by the distances at which the presser foot can rest flat on either side of the seam.Source: §5 · p. 6, Figure 9 · p. 6
6 Implementation
StitchOver is a standalone desktop application that imports and exports SVG stitch plans. Its canvas lets users align designs and annotate seams with editable properties.
6 Implementation
- StitchOver is a standalone Python and PySide6 desktop application that converts SVG stitch plans and exports SVG for the embroidery-machine digitizer.Source: §6 · p. 6
- The millimeter-based canvas supports pan, zoom, direct manipulation, and optional fabric photographs for visual alignment.Source: §6 · p. 6
- Users annotate straight or curved seams and edit seam width, buffer distances, fold direction, and per-region stitch parameters.Seams automatically partition the fabric into regions.Source: §6 · p. 6
7 Technical evaluation
The technical evaluation examines embroidery quality across plain fabric, seamed fabric, and seamed fabric with JumpStitches. It uses conductive thread and measures defects, resistance, and positional error.
7 Technical evaluation
- Each evaluation compares plain fabric, seamed fabric, and seamed fabric with JumpStitches, measuring defects, conductive-thread resistance, and positional error.Source: §7 · p. 6
7.1 Materials
- Experiments used a serviced Brother PE800 at its lowest speed with Madeira HC 40 conductive top thread, polyester bobbin thread, and a universal embroidery needle.The substrate was PU leather with 6 mm overlocked seams; the thread tension dial was set to 4.Source: §7.1 · pp. 6–7, Figure 10 · p. 7
- The standard pattern was 28 parallel 90 mm rows of 4 mm running stitches, repeated across 10 samples.A connected-grid capacitive proximity sensor provided a lower bound on pattern complexity, with a more complex pattern used later.Source: §7.1 · p. 7
7.2 Metrics
- Visual inspection classified defects as minor aesthetic issues, functional defects, or process failures.Minor defects may accumulate and precede more severe failures during prolonged embroidery runs.Source: §7.2 · p. 7
- Resistance was measured on one visually defect-free trace per sample because visible mechanical failures can break continuity and make resistance measurements uninformative.The conductive coating can be damaged under suboptimal conditions even when a trace has no visible defect.Source: §7.2 · p. 7
7.3 Simple seams: JumpStitches eliminate defects from 70% (baseline)
JumpStitches eliminated defects in the simple-seam test and largely overcame the effects of belt-tension variation. Without JumpStitches, seam defects became more frequent and varied as tension changed.
7.3.1 Results.
- 7/10 seamed samples had defects without JumpStitches, versus zero with them; JumpStitches also largely overcame tension-related problems.The simple-seam comparison included plain fabric, seamed fabric with traditional digitization, and seamed fabric with JumpStitches.Source: Figure 11 · p. 7, §7.3.1 · p. 8, §7.4.2 · p. 8
7.4 Belt tension: variations of belt tension increase defects, but not with JumpStitches
- With JumpStitches, only 2 over-tensioned and 1 under-tensioned samples had defects, which were primarily aesthetic.The remaining defect types included bird nesting, floated stitches, and supporting-thread penetration.Source: §7.4.1 · p. 8, §7.4.2 · p. 8
7.4.1 Results.
- Over-tension caused defects in 9/10 seam samples, including thread breakage, fraying, bird nesting, and pattern distortion; under-tension caused 7/10 defects, mainly floated stitches.Under-tension shifted the dominant defect type from looping stitches in the well-calibrated condition to floated stitches.Source: §7.4.1 · p. 8, Figure 15 · p. 9
- Resistance showed no significant difference with JumpStitches, while over-tension increased resistance variability in the seam condition.The authors interpret the high resistance variance as suggesting that stitches may be electrically compromised without being visibly broken; JumpStitches reduced this variability.Source: §7.4.1 · p. 8, §7.4.2 · p. 8
7.4.2 Conclusion.
- The remaining JumpStitch defects were aesthetic, and such defects also appeared in the plain condition.Source: §7.4.2 · p. 8
- Under over-tension, defects appeared across the pattern rather than only near seams; under well-calibrated and under-tension conditions, failures were localized around seams.The authors attribute the broader over-tension behavior to the uneven surface introduced by seams destabilizing already high tension.Source: §7.4.2 · p. 9, Figure 16 · p. 9
7.5 Pattern orientation does not impact the results
Pattern orientation did not significantly affect defect rates, while JumpStitches improved seam-crossing results under three-way seam stress but did not eliminate every defect.
7.5.1 Results.
- The orientation test compared patterns crossing seams perpendicularly while aligned differently to the machine axes.The remaining JumpStitches defects were support penetration and floated stitches, associated with tension miscalibration rather than orientation.Source: §7.5 · p. 9, §7.5.2 · p. 9
7.5 Pattern orientation does not impact the results
- 0/10 horizontal and 2/10 vertical JumpStitches samples were defective, versus 7/10 and 5/10 with traditional digitization.Rotating the pattern produced no meaningful change in defect rates, and resistance means were comparable across orientations for all three construction methods.Source: §7.5.1 · p. 9, Figure 17 · p. 9
7.6 Complex seams require slight modifications
- At three-way seam intersections, JumpStitches alone did not fully solve the problem, so the tool offers global rotation or slight needle-point perturbation.The evaluation in this section uses the rotation solution; which modification is desirable can depend on circuit functionality.Source: §7.6 · p. 9
7.6.1 Results.
- For three-way seams, resistance averaged 72.9±11.9 Ω with JumpStitches and 78.9±14.6 Ω without StitchOver.Both conditions had higher resistance variability than the single-seam Plain baseline of 83.4 ± 3.6 Ω.Source: Figure 18 · p. 9, §7.6.1 · p. 10
- 100% to 40%: on three-way seams, JumpStitches reduced defects from 10/10 samples to 4/10.The without-StitchOver condition had thread fraying in 10 samples, thread breakage in 10, and bird nesting in 7; JumpStitches had fewer affected samples across defect categories.Source: §7.6.1 · p. 9, Figure 18 · p. 9, §7.6.2 · p. 10
8 Discussion
JumpStitches reduce or often eliminate defects across common simple seams, while more complex seams and material variation introduce limits that call for adaptations.
8.1 Effectiveness of JumpStitches
- JumpStitches significantly reduce and often fully eliminate defects across common simple seams; greater seam or pattern complexity can place crucial needle points on seams.The tool suggests global rotation or needle-point perturbation, leaving the choice to users because modifications may affect functionality differently.Source: §8.1 · p. 10
- Defect types differ across conditions, potentially helping diagnose gradual machine decalibration or poor tuning to different materials.The authors describe this as an insight for debugging by visually inspecting results, not as a core contribution.Source: §8.1 · p. 10
8.2 Embroidery on non-planar textile structures
- Non-planar regions can be divided into nearly flat patches and embroidered separately, while tubular regions may use free-arm embroidery or floating hooping.Feasibility for tubular structures depends on access to a single-layer embroidery area and the machine setup.Source: §8.2 · p. 10
8.3 Material considerations
- The material evaluation focuses on thick leather and does not systematically characterize parameters across diverse fabric combinations.Exploratory tests indicate that leather is more prone to thread breakage and woven fabrics to pattern distortion, with these issues more pronounced on seams.Source: §8.3 · p. 10
9 Application Opportunities on Textile Goods
The application discussion describes seam-sensitive garment placements and demonstrates shoulder electrodes embroidered across a Y-shaped seam using separate patches.
9 Application Opportunities on Textile Goods
- Shoulder, cuff, joint, RFID, and underarm applications face placement challenges where seams, folds, or curvature interrupt technical embroidery.The examples include capacitive touch, stretch sensing, antennas, and sweating-sensing structures.Source: §9 · p. 10
- A cotton-shirt demonstration placed capacitive electrodes across a Y-shaped shoulder seam using two locally flat patches and floating hooping.Automatically generated JointStitches electrically connected the conductive traces on the separately embroidered patches.Source: §9 · pp. 10–11, Figure 20 · p. 11
10 Conclusion
The conclusion presents JumpStitches and StitchOver as a way to reduce fabrication defects when embroidering technical patterns across seams, supporting applications on everyday textiles.
10 Conclusion
- The technical evaluation found a vast decrease, and mostly elimination, of fabrication defects caused by seams using JumpStitches.StitchOver supports inserting the stitches into existing embroidery patterns and is designed to address subtle belt-tension variations.Source: §10 · p. 11
- The authors present the work as a step toward broader technical-embroidery use cases on everyday fabrics and garments.They state that the software is being distributed as an open-source package for community extension.Source: §10 · p. 11