An ecommerce company is automating the categorization of its products based on images. A data scientist has trained a computer vision model using the Amazon SageMaker image classification algorithm. T

Question

Options

• AUse the SageMaker semantic segmentation algorithm to train a new model to achieve
• BUse the Amazon Rekognition DetectLabels API to classify the products in the dataset.
• CAugment the images in the dataset. Use open source libraries to crop, resize, flip, rotate,
• DUse a SageMaker notebook to implement the normalization of pixels and scaling of the
• EUse Amazon Rekognition Custom Labels to train a new model.
• FCheck whether there are class imbalances in the product categories, and apply

• A
  6% (1)
• B
  78% (14)
• D
  11% (2)
• F
  6% (1)

Explanation

Three complementary strategies improve accuracy here. Option C (data augmentation) artificially increases dataset diversity by applying transformations like cropping, flipping, and rotation, helping the model generalize to new products. Option E (Amazon Rekognition Custom Labels) provides a managed service purpose-built for custom image classification, typically requiring fewer labeled examples to reach production-quality accuracy quickly - aligning with the 'as soon as possible' requirement. Option B (Amazon Rekognition DetectLabels) leverages a pre-trained model to enrich the existing dataset with additional label context, supplementing the custom training. Option A (semantic segmentation) is designed for pixel-level tasks, not product categorization. Option D (normalization) is a minor preprocessing step that alone won't dramatically boost accuracy. Option F (class imbalance correction) is a valid concern but not among the top-three highest-impact steps compared to the chosen options.

No community discussion yet for this question.