Object Recognition Techniques
-
Conditions
-
Object recognition is studied by observing a person's perceptions and recognition of objects with various conditions. Changing these conditions often does not change a person's recognition of an object. Object recognition proves that objects can be different sizes, shapes and colors and a person still recognizes them. Along with that, the orientation of an object does not matter either. Objects sideways, upside-down and backward often are still recognized by people. The lighting and background behind an object is also a technique used in studying object recognition. 2-D and 3-D images as well are still likely to be recognizable by people.
Recognition by Components
-
Recognition by components (RBC) is a technique that is used for object recognition. If only a few components of an object are shown to a person, will the person still identify the object? In most cases, this has proved to be true, as long as the object is a familiar item. With this same topic, objects surrounded by clutter are also studied to see if clutter makes a difference on object recognition.
Robots
-
Object recognition has been tested with robots. At the 5th Asian Conference on Computer Vision, an abstract was written based on the findings of robots. Robots performing work require certain environments for them to recognize their job. Through the study of these robots, 2-D and 3-D images were placed in the robot's database. This technique enabled the robots to identify the images they needed to understand. Once the robots understood and recognized the objects, they produced their work easily and more rapidly. In order for robots to complete complicated jobs, the development of advanced object recognition is crucial.
-