A fabric defect is cheap when you find it and expensive when the buyer does. Caught at inspection, a flaw means a re-grade, a cut, or a small allowance. Caught after the fabric has been cut into garments and shipped, the same flaw means rejected goods, a chargeback, and a dent in a buyer relationship that took years to build. That asymmetry is why fabric inspection exists — and why the way defects are detected is quietly one of the highest-leverage processes in a textile business. This guide covers the common fabric defects, what causes them, and how AI is changing detection.
Common fabric defects and their causes
Most fabric defects trace back to a stage of production. **Weaving defects** include broken ends and picks (missing warp or weft yarns), floats, and reed marks — usually from yarn breakage or loom issues. **Yarn defects** such as slubs, thick-and-thin places and neps show up as irregularities in the cloth and trace back to spinning. **Dyeing and finishing defects** include shade variation, dye-lot mismatch, streaks, patches and crease marks — often the most costly because they affect whole rolls. **Handling defects** like holes, stains, oil marks and selvedge damage come from mechanical mishandling anywhere along the line.
Knowing the cause matters because inspection is not only about rejecting bad fabric; it is about feeding the cause back upstream. A recurring reed mark is a loom to fix; a recurring shade variation is a dyeing process to control. Inspection that only grades, without capturing the defect type and its source, throws away half its value.
The 4-point inspection system
The industry standard for grading fabric is the 4-point system, which assigns penalty points to defects by size — a small defect scores one point, a large one up to four — and totals the points per 100 square yards to decide whether a roll passes. It is objective, widely accepted by buyers, and turns a subjective 'this looks bad' into a defensible number. Its weakness is that it depends entirely on a human inspector seeing every defect as fabric runs past on a light table, hour after hour — and human attention is not constant.
The limits of manual inspection
Manual fabric inspection is skilled work, but it fights biology. Fabric runs past at speed; a tired inspector at the end of a shift misses what a fresh one catches; and two inspectors grade the same roll differently. Studies of visual inspection across industries consistently find that human detection of small defects at speed tops out well below 100%. The result is that some defects always slip through — and which ones slip through varies with who is inspecting and when. That inconsistency is the real cost, because it means a buyer occasionally receives what your own system said had passed.
How AI computer vision detects defects
This is where computer vision is genuinely changing the process. An AI vision system watches the fabric through cameras as it runs, and flags anomalies — holes, stains, weaving faults, shade variation — in real time, at a consistency a human cannot sustain across a shift. It does not get tired, it grades every metre the same way, and it can catch a developing defect (a slowly worsening streak) earlier than a person scanning for discrete flaws. Paired with the 4-point framework, it does not replace the standard; it applies it consistently. Read honestly: AI inspection is not magic and needs training on real defects, but on the specific job of spotting the same defect the same way on the thousandth metre as on the first, it beats human attention decisively.
Why detection is only half the value — capturing it is the other half
Finding a defect is worthless if the finding is not recorded. The real return on inspection comes when every defect is captured as data — type, size, position, roll, lot, machine — so patterns become visible. That is what turns inspection from a gate into a feedback loop: the loom that keeps producing reed marks, the dyeing batch that keeps varying, the supplier whose fabric keeps arriving with slubs. A quality control module records defects against rolls and lots, and the garment ERP and wider textile platform tie the finding back to the stage that caused it — so the same defect stops recurring.
AI and 4-point and a light table are all just ways of finding the flaw. What makes inspection pay is finding it before the buyer does, and using what you find to stop it happening again. If defects are caught inconsistently and their causes are never traced, that is where quality — and margin — leaks. You can see how defect data flows back upstream with your own fabric and machines.
Frequently Asked Questions
What are the most common fabric defects?
Common defects group by cause: weaving defects (broken ends and picks, floats, reed marks), yarn defects (slubs, thick-and-thin, neps), dyeing and finishing defects (shade variation, dye-lot mismatch, streaks, patches, crease marks), and handling defects (holes, stains, oil marks, selvedge damage). Capturing the defect type helps trace the cause back upstream.
What is the 4-point system for fabric inspection?
The 4-point system grades fabric by assigning penalty points to defects by size — one point for small defects up to four for large ones — and totalling points per 100 square yards to decide whether a roll passes. It is the industry-standard, buyer-accepted method that turns subjective judgement into a defensible number, but it depends on a human inspector seeing every defect.
How does AI detect fabric defects?
An AI computer-vision system watches fabric through cameras as it runs and flags anomalies — holes, stains, weaving faults, shade variation — in real time, grading every metre consistently without fatigue. It applies the 4-point standard uniformly and can catch developing defects earlier than a person. It needs training on real defects but beats human attention on sustained, consistent detection.
Vastra ERP Editorial Team
Textile Technology Experts
Our editorial team brings decades of combined experience in textile manufacturing, supply chain management, and enterprise technology. We publish in-depth guides, industry analysis, and practical insights for textile professionals worldwide.
