If it did not, it could be because it needs more and different kinds of data belonging to the category you choose. According to you, what kind of data will be needed by the AI model to recognise the image correctly?
The AI model needs diverse training data including images taken from different angles, in various lighting conditions (bright light, shade, indoor lighting), with different backgrounds, showing partial and complete views of the object, and captured at different stages (if applicable). For example, to recognize a banyan tree, the model needs images of the tree in shade and light, at different stages of growth including saplings, from different angles, in different geographies, and both full and partial photos. More variety in training data helps the AI model recognize images it has never seen before.
Explanation
The chapter emphasizes that to train a machine to recognize images effectively, one must upload many different images showing the object in various conditions. Aisha's example shows she collected 12-15 different images for each mango variety to ensure sufficient data for the AI to learn better.