Recognition might begin with the identification of features in the input pattern. With these features appropriately catalogued, you can start assembling the larger units. The features that are relevant here are not the features of the raw input. Instead, the features that we use are the ones in our organized perception of the input. We recognize objects by detecting the presence of the relevant features.
Advantages of a feature-based system:

1. Features such as line segments and curves could serve as general-purpose building blocks. Not only would these features serve as the basis for recognizing letters, but they could also serve as the basis for recognizing other, more complex visual patterns, opening the possibility of a single object-recognition system able to deal with a wide variety of targets.
2. We have noted that people can recognize many variations on the objects they encounter - A's in different fonts or different handwritings. But although the various As are different from each other in overall shape, they do have a great deal in common: two inwardly sloping lines and a horizontal crossbar. Focusing on features might allow us to recognize As despite their apparent diversity.
3. Several lines of data indicate that features do have priority in our perception of the world. The difference between angles and cuves jumps out immediately. Same holds for differences in color, orientation, and many other simple features.
4. Other results suggest that the detection of features is a separate step in object recognition, followed by subsequent steps in which the features are assembled into more complex wholes.