Marrs theory and a complete account of perception
Published: Last Edited:
Disclaimer: This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
Any opinions, findings, conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of UK Essays.
Visual perception plays great importance in peoples everyday lives. It is an interpretation of what humans and animals take in trough their senses. Perception helps us to interact with people, to play a sport, to judge a situation and mostly having accurate visual perception helps humans and animals to survive by being able to react to outside stimuli. Perception happens so effortlessly that it is hard to believe the complexity behind it (Eysenck and Keane, 2005). Different theories of perception have been proposed, while some of them look at perception as object recognition others look at perception as a need for action. This essay will look at whether Marr's (1982) theory, Biederman's (1987) and Riddoch & Humphreys's (2001) theory, provide both a valid and a complete account of perception. To do that one must look more specifically at these theories and how they explain perception. As well one must consider the strengths and limitations of these theories and look at other possible explanations of perception what will lead to final conclusion on validity and compliance of the theories of perception.
First we will look at description of Marr's theory, which is very huge, influential computational models, build for visual perception. He used neuropsychology and experimental psychology to support his theory (Eysenck and Keane, 2005). When input (the object we looking at) hits the eye it consists simply of variations of brightness of light and colour of light. Visual perception takes the light pattern and extracts from it the shapes and forms that are in the world. The processes of object recognition are done in four stages. Each stage takes the output from previous stage, processes it and hands it to the next stage. The final output is three dimensional representation of the object that was perceived from the world. This representation is used to search the memory to find a match, once the match is found the object is recognised (Eysenck and Keane, 2005). Marr's (1982; cited in Eysenck and Keane, 2005) first three stages are about analysing this pattern of light and pulling out the shapes that automatically form what we see. The fourth and fifth stage is about depth. Adding depth to shapes, helps for example, view face as something with additional information such as eyes are deeper in face, ears are set and so on. Not as a flat object. Thus, Marr tried to explain perception of object recognition trough different stages.
The second is Biederman's recognition by components theory, which is theory of human experimental psychology. He developed his theory doing experiments with human beings. He used the foundation of Marr's theory to build his own. His theory kind of fills in the gaps of Marr's theory. Biederman's theory is more about recognition then perception, is on motion of geons, which are geometric objects. This theory argues, that most objects are composed of the same kind of shapes put together. E.g. human beings can be seen as 4cilinders, 2 for arms and two for legs, head can be seen as a blob. This kind of geons are seen everywhere, in almost every object (Beiderman, 1987). It is believed that there is around 36 basic geons that are used to form objects. Biederman in his theory argues that humans extracting three dimensional representation of the world is all about identifying the geons which the perceived object is made of. Once the component geons which make the object are identified, matching with objects stored in memory starts. Once the match is found object is recognised (Eysenck and Keane, 2005). As Marr's theory, Biederman explains his theory trough stages.
Riddoch and Humphereys' (2001) (R&H) theory of perception and object recognition is built on looking at patients with brain injury. They use evidence from neuro cases such as people with stroke, agnosia and different visual impairments and put it together with Marr fundamental theory to build their own theory. Like a Marr's theory, it has stages of perception. However their stages are backed up by neuropsychological evidence. Accordind to Marr and Riddoch & Humphreys at one of the first stages of visual perception there is a point where the sketch is being composed (Eysenck and Keane, 2005). This means that the perceptual system is assembling the perceptual information about lines, edges, where they start, where they end and where they meet with another line or edge and so on. Riddoch and Humphrey found a neurological case that could explain and support this stage. Person known as DF could not do basic perceptual things, like know where one edge begins and where one ends. She could not tell how corners from one object can be distinguished from corner of different object (Milner et.al. 1991). Riddoch and Humphey saw it as an evidence of important stage in visual perception. The other case that they took as an evidence of stages in perceptual analyses, was a patient known as HJA who could not pick up different components of an input, he lacked the ability of putting different components together for example face components. He could distinguish ear, nose, lips and eyes but he could not put it together as a face (Riddoch&Humphrey 2001, cited in Eysenck and Keane, 2005). They used a number of different cases which supported with neurological evidence the stages of perception.
The three models discussed above describe very important and relevant ideas about perception. However, it is important to look at the supporting and conflicting evidence for all three arguments in order to understand the validity of these theories.
One of the tasks that support Marr's and Biederman's hypothesis is where participants had to decide if two line drawings matched each other or were different when it comes to the length of the lines drawn. When the main axes were shortened it influence the performance of participants. (Lawson and Humphreys, 1996; cited in Eysenck and Keane, 2005). This task supports the hypothesis of detecting major axis in object recognition. Another support comes from Hummel and Biederman's (1992; cited in Eysenck and Keane, 2005) with their studies on connectionist model. Participant was trained to recognise ten different objects from one point of view, after which objects could be recognise from any point of view. When objects presented in different orientation the recognition got worse. It has proved to be very sufficient model for explaining the binding problem. More support comes from Cooper and Biederman (1993; cited in Eysenck and Keane, 2005) when it comes to explaining the importance of geon detection. Further evidence comes from experiment done on monkeys, where Vogels, Biederman, Bar and Lorincz (2001; cited in Eysenck and Keane, 2005) suggested existence of geon-sensitive neurons in monkeys. Nevertheless, cautious has to be applied when it comes to taking into account these pieces of evidence. As one may have noticed, Biedermans in a number of these researches played a role what may have led to influence of results. Fery and Morais (2003) reported a patient who was very good at tasks that involved recognising and matching shapes, but was able to identify only 16% of these objects. This case support Riddoch & Humphereys' theory. Another support comes from study done by Hillis and Caramazza (1991) on patient with category specific deficits. This provides the ability to recognise living things while having the inability to recognise non-living things. Further support comes from study done by Humphreys and Riddoch (1987) on patient HJA which suggests a deficiency in feature binding and shape segmentation. Miller et.al (1991) looked into damage to the "grouping by colinerarity" which results support Riddoch and Humphereys theory. Finally, neuropsychological patients with difficulties to recognise objects from unusual views indicate the ability to rotate a three dimensional representation is a key feature of object recognition (Warrington and Taylor; 1978; cited in Eysenck and Keane, 2005). This theory is strongly supported with neuropsychological data, which will help it improve and define in the future.
However, to evaluate and to understand the validity and complicity of these three theories of object recognition and perception one must look at the contrasting evidence. Marr's and Biederman theories are limited when it comes to recognition within category. E.g. when it comes to faces these theories would have difficulties to distinguish different faces within a group. They provide limited explanation for subtle perceptual distinction within categories (Eysenck & Keane 2005). In these theories object recognition happens trough matching an object-centred representation independently from the observer's view, with the objects stored in memory. However, there is great evidence for view point dependent recognition of an object. What indicates oversimplification of their theories (e.g., Gauthier & Tarr, 2002; Tarr&Bulthoff, 1995 cited in Eysenk & Keane 2005). Palmer (1975; cited in Eysenk and Keane, 2005) proposed how important is context when it comes to influencing object recognition. Recognition is easier if object is harmonious with the environment, if there is no context recognition is average, and much more difficult if there is no consistency. One of the key evidence against Biederman and Marr's theory is the fact that many things in the world do not have a geon structure. e.g. sky, sea (Eysenck and Keane, 2005). Colour and colour knowledge has been shown to play an important part in object recognition. This finding by Joseph and Proffitt (1996; cited in Eysenck and Keane, 2005) does not support Biederman's theory in which the edge information is important for object recognition.
Moreover, these theories do not consider the importance of environment on perception of an object. When it comes to Riddoch and Humphreys' theory, even though it certainly explains well visual impairments, there are still some limitations. Firstly, the theory should be seen as a framework rather than complete theory (Eysenck and Keane, 2005), since the stages using output from previous stage is not clearly explained. Secondly, the assumption is that recognition happens trough bottom-up way. However, one should not dismiss the fact that it is likely that there are some top-down processes during object recognition (Eysenck and Keane, 2005). These contradicting evidence suppresses validity and compliance of these three theories above mentioned.
Different view on perception comes from Gibson theory which almost ignores the function of perception. This theory is concern with the extent to which we use perception for action, for moving around in the world and handling things. Gibson approaches perception from different angle. His theory is called ecological direct theory of perception (Gibson 1950, 1966, 1979; cited in Eysench and Keane, 2005). It is called ecological because he is not interested how single object is perceived but how the whole world is perceived, the whole visual input. His theory is direct because it doesn't involve processing stages as do the above mentioned theories of object recognition. According to him all the information required for perception is present in optic array. The optic array is the pattern of light that reaches the brain while perceiving object. Optic array is picking up invariants which even when on the move the invariants stay the same e.g. texture gradient (Eysenck and Keane, 2005). The only thing that gets Gibson close to talking about object recognition is his idea of affordance. The affordance of an object is the function it offers its self for e.g. being seated on, being grasped. This theory definitely sees action as the importance part in perception, rather than as recognition as purpose for perception as the aforementioned theories did.
Another perspective that should be taken into consideration when it comes to looking at perception and its validity is the argument that there are two separate pathways in the brain that deal with different types of information for different purposes. This has initially been proposed by Ungerleider & Mishkin (1982; cited in Goodale and Millner, 1994). This theory argues that the bottom (ventral) pathway is responsible for object recognition, known as "what pathway". If this pathway is impaired in some way then the person looses the ability to recognise the object. The top (dorsal) pathway is responsible for determining the spatial location of an object, known as "how" pathway (Eysenck and Keane, 2005). If this pathway is impaired in some way, the person can not do active things e.g. cannot pick things up. Data from electrophysiological, to anatomical and behavioural records verified the existence of these two pathways. There are these two different pathways in the brain, one seems to be concern with perception for recognition while the other seems to concern with perception for action. This perspective in short offers explanation on object perception for recognition and object perception for action as well as it adds valuable information to understand perception.
To summarise, Marr's (1982), Biederman (1987) and Riddoch & Humphreys (2001) theories added great value to our understanding of perception. However, there is still considerable number of evidence that weakens these theories. Where Marr, Biaderman and R&H defended object recognition, others pointed out the weakness of their theories. One of the major consideration to be taken into account is that the supporting evidence in large is produced by research groups that included the theorists, so they may be considerably biased in designing the experiments, or in their interpretation. Additionaly Gibson's argument for perception for action introduced new view of perception. However, the theory that is the closes to understanding perception is the dual pathway proposition, in which each pathway uses perception for different purposes. Therefore to conclude we can say that Marr's (1982), Biederman (1987) and Riddoch & Humphreys (2001) do not provide valid and complete explanation for perception, however these theories provided strong ground for building on theories on perception such as the dual pathway, which is the closes to explaining perception.