Skip to main content
Example · AI & machine learning

Defect detection with scarce labelled data

Cormorant needed to detect rare manufacturing defects from camera images, but had only a few hundred labelled examples of each defect class. Off-the-shelf models and standard augmentation didn't generalize.

Cormorant Vision (fictional). This fictional example is for education only. It is not a customer claim or tax advice. It shows how to structure and reason about a SR&ED narrative. Do not copy it into a claim.

Technological uncertainty (T661 line 242)

Weak

We used machine learning to build a defect-detection model, which was challenging with limited data.

Strong

It was uncertain whether acceptable detection accuracy on defect classes with fewer than 300 labelled samples was achievable. Standard transfer learning and augmentation had been tried and produced high false-negative rates, and it was unknown whether a synthetic-data or few-shot approach could close the gap without introducing artifacts the model would learn.

The weak version says 'ML is hard'. The strong version pins the uncertainty to a measurable condition (few-shot, <300 samples) and states what was already ruled out.

Technological advancement (T661 line 246)

Weak

We built a model that detects defects with good accuracy.

Strong

We advanced our understanding of few-shot defect detection: a physics-informed synthetic-augmentation pipeline combined with a contrastive pre-training step reduced false negatives materially in controlled tests. We also established that naïve GAN-generated samples degraded precision, which redirected the approach.

The strong version reports the knowledge advance and a specific negative result, not a product milestone.

Evidence matrix

What could support a claim like this

The matrix pairs each described element with the kind of source that could support it. The actual records still need review.

SourceWhat it shows
Experiment logs & metricsThe systematic runs, hypotheses tested, and per-class accuracy over time
NotebooksThe augmentation and pre-training experiments, including the discarded GAN approach
Model-version historyWhich changes moved the metric, tying effort to result
Commit historyThe dated iteration behind the experiments

The takeaway

Describe the investigation, not the model as a finished product. Show the baseline, competing approaches and negative results, then review that work against both CRA requirements.

All examples · SR&ED for AI & machine learning

Draft yours from real evidence

Use approved GitHub activity, uploaded documents and project records to prepare an editable narrative draft for review.

7-day free trial · No credit card required