##### CASE STUDY

# Digital Watches

**A globally-renowned electronics company sought our expertise to elevate their defect detection capabilities within their digital watches, analyzing for defects on the watch faces.**

**Mindtrace was selected to create an industry-leading AI model that seamlessly integrates into their existing model backbone, reducing disruption while revolutionizing their manual quality assessment procedures.**

**Challenges**

- Under 300 available training data images
- Unlabelled training data
- Images contain high levels of noise

### Step 2 - Model Testing

After the training phase was complete, the model was tested on further data samples, achieving a 97% defect detection rate.

### Step 1 - Model Training

During the training phase, Mindtrace's Brain-Sense™ Platform successfully processed 300 training images utilizing few-shot learning due to the limited information present in the dataset.

### Step 2 - Model Testing

After the training phase was complete, the model was tested on further data samples, achieving a 97% defect detection rate.

### Step 1 - Model Training

## Outcome

**Despite the constraints of a limited dataset and the presence of noise, Mindtrace’s Brain-Sense™ platform effectively developed a bespoke AI solution that achieved a 97% defect detection rate within the training data.**

**Furthermore, this solution seamlessly integrates into their existing model backbone, with deployment options available for on-premises, cloud, or edge devices, ensuring minimal disruption.**

**AI Brains Used**

1

##### Detection

Identifies defect types, severity & location.

1

##### Classification

Object identification & classification.

1

##### Analytics

Generates insights & recommendations.

[Learn more](/content/detection/index.html)

[Learn more](/content/classification/index.html)

[Learn more](/content/analytics/index.html)

1
Learning Samples
1%
Defects Detected
1
Delivery Time