Which steps would improve the accuracy of the solution?
(Choose three.)
Use the SageMaker semantic segmentation algorithm to train a new model to achieve improved accuracy.
Use the Amazon Rekognition DetectLabels API to classify the products in the dataset.
Augment the images in the dataset. Use open source libraries to crop, resize, flip, rotate, and adjust the brightness and contrast of the images.
Use a SageMaker notebook to implement the normalization of pixels and scaling of the images. Store the new dataset in Amazon S3.
Use Amazon Rekognition Custom Labels to train a new model.
Check whether there are class imbalances in the product categories, and apply oversampling or undersampling as required. Store the new dataset in Amazon S3.
Explanations:
The semantic segmentation algorithm is unsuitable for image classification, as it focuses on segmenting objects within an image rather than classifying images as a whole.
Amazon Rekognition’s DetectLabels API is not designed for fine-tuned classification of specific products and may lack the customization needed for the company’s unique product categories.
Image augmentation can help the model generalize better by exposing it to a variety of transformations, which improves its performance on new, unseen data.
Normalizing and scaling images help standardize the input data, allowing the model to better recognize patterns and improve accuracy.
Amazon Rekognition Custom Labels could help with classification, but retraining the existing SageMaker model using enhancements like data preprocessing and augmentation would be more efficient.
Addressing class imbalance through oversampling or undersampling helps the model learn to classify underrepresented classes, which improves overall accuracy.
I opine that the answer is:
Augment the images in the dataset. Use open source libraries to crop, resize, flip, rotate, and adjust the brightness and contrast of the images.