You have a labeled image dataset with two classes: lion or cheetah. You want to train a model that, given a new image, predicts whether it contains a lion or a cheetah. Which SageMaker algorithm choice is appropriate?

Image classification using TensorFlow is the correct choice because the problem requires classifying an entire image into one of two predefined categories (lion or cheetah). This is a fundamental task in image classification. XGBoost is a tabular data algorithm, not suitable for image data. Object detection would identify and localize multiple objects within an image, which is more complex than needed here. Semantic segmentation would classify each pixel in an image, assigning it to a specific class, which is also overkill for a simple whole-image classification task.