An Object model Of recognition we will kick at is Mar and Insignia’s special process used to generate an object centered D description, ND a facial recognition model we will compare this too will be the Connectionist model in which there is a structure which taps into the next resource of information available to the viewer. Recognizing faces and objects are considered separate processes and there has been debate over which process this is, if its innate or if its learned.
The debate stems from the question as to whether recognizing faces is a special process and takes separate neurological pathways than that of recognizing an object. Theories of expertise recognition suggest that we have a specified skill in recognizing aces which are different from the recognition of objects due to the exposure and social deterrents that faces have on us, using people in a field Of another skilled recognition, e. G. Dog show judges to test if there is a difference between faces and just other expertise objects.
In objection of this idea is the theory of Domain specificity, which indicates that the reason faces are processed differently is because of the different neural structure the brain uses to process a face than that of an object, using fem. scan from Prognosis (the inability to recognize faces) patients and Copperas delusion patients who do not recognize objects) to further their argument that the process happens in two different regions of the brain, not just because we are more skilled at facial recognition, but that facial recognition is a special process.
Mar and Insignia suggested three-dimensional objects are recognized by breaking them up into generalized cones, you can do this by observing the ‘occluding contours’ (the objects silhouette). This analysis allows any object to be described within a canonical coordinate frame (the process which involves the description of all standard objects). There is 3 steps to this process, 1 .
Derive the object shape by the following three assumptions, (a) each point on the silhouette matches only one point on the D object, (b) points near each other in the D image are near each other in the D object and (c), all points on the silhouette lie in the same plane. If any of these assumptions do not match, the object may not be recognized properly. Step number 2, Locate the objects component axis and derive a D description, you do this by dividing the object into component parts by linking areas of concavity, you then find the axis for each of these dividends and then link all together to form the D ascription.
Step 3 is the comparing this D description to a mental catalogue of D models all previously seen and stored, the more descriptive matched the more detail at each level surfaces, when the match is found the process stops and the object is recognized. Studies to support Mar and Insignia’s module include Lawson and Humphreys (1996) experiment in which patients were shown line drawings in where the major axis was rotated, making it difficult to locate, the results showed that recognition was negatively affected.
Arguments against this model suggest that if all conversions of an object can e generalized to cones then all exemplars of the cones would be assigned to the same category, meaning we would not be able to tell the difference between in-been-categories, also a fall back to this theory is the extensive capacity for a stored catalogue of images a person must have for comparison.
Similar to Mar and Insignia’s model of recognition is The Connectionist model of facial recognition, this model also uses the idea that complete analysis happens in stages involving more information at each level, it derives from a simpler older model referred to as a cognitive theoretical framework y Hay and Young (1982) Young et al (1985) and Bruce and Young (1986).
Burton and Bruce (1993) developed the IAC Interactive activation and competition network) model, the model compromises of units which are organized into pools, the pools consist of Frog’s, there is one FUR for every familiar person seen which then leads to the activation of the PINS, this is where a face is classified as belonging to a certain person and confirming the familiarity of that person, if the PIN is activated then it will lead to ISIS where semantic information of the person is stored (occupation, nationality, etc. ND then finally onto Lexical output, which would be for example, their name. These links are bidirectional and excitation can also run up the way from PIN to FRI. also. Like the above object recognition model, any error in these stages can lead to mistaken recognition; Young et al (1985) conducted a study to further the legitimacy of the theory by using 22 participants to record their recognition mistakes in a diary over the course of 8 weeks, 3 groups of errors were found, 1.
Mistaking someone familiar as unfamiliar and vice versa, 2. Being unable to ‘place’ someone and 3. Recognizing someone but rejecting it cause of doubt. The model can be used to explain a number of different phenomena in the everyday process of recognizing people, and the findings of the study itself are consistent with the notion that the recognition of faces involves a sequence Of processes using different types Of information.
Propagandist is the inability to recognize faces whilst maintaining the ability to recognize other objects, Young et al (1993) conducted a study to show that although conscious awareness of the face if affected, unconscious awareness may not and that recognizing someone’s face may be an entirely different recess to emotion or expression detection, in their study they found the participants with unilateral focal lesions displayed different types of dysfunction, their findings showed that participants had the ability to recognize familiar faces, ability to match unfamiliar faces and identifying facial expression occurred independently.
The results suggested that facial recognition may be overt (conscious) being aware of recognition or covert (unconscious) responding without knowledge of who the person is. Another phenomenon termed the Copperas delusion is the belief that a person lives someone they know, usually familiar like a relative or a close friend, has been replaced by an imposter, this is because they recognize the face but they simultaneously refute its authenticity, and do not convey the same emotion they would for the ‘real person’.
The expertise model suggests that the IAC model can explain these phenomenon by the weakening of the connections between Frog’s and PIN’s, so that the next Step is not high enough for conscious awareness. Theorists of Domain specificity argue that if facial recognition can be affected separately from object recognition it clearly implies that the neural pathways for recognizing faces are special compared to those of objects and take different neural processes to work.
One of the hallmarks of face processing is that faces are processed Holistically, also called configured processing, seeing the face as a whole rather than an array of separate parts. Studies into what is known as the Inversion effect support the domain specificity debate and the configured processing idea, showing when a researcher inverts a photograph it impairs recognition more so than inverting a picture of an object, e. . A table or a boat. Yin (1969) who’s inversion studies showed that we have better memory for faces than other objects and that inverting a face disrupts it recognition more than inverting an object by testing participants with stimulus material such as house and airplanes.
Diamond and Carrey (1986) tested the alternative hypothesis using the expertise model to test the inversion effect and found when they tested dog-experts, the were just as likely to be affected by the inversion of the dog than human faces, they interpreted these findings in that face recognition is not a separate process to object recognition, but that we evolve special expertise at recognizing faces, they explained this happens by using First order properties, that help us to recognize it as a face by calculating spatial relationships between eyes, nose and mouth and Second order relational properties, helping us to say who belongs to this face being by being able to calculate how far apart features are from each other.
McKeon et al retested Diamond and Caress hypothesis and failed to replicate the original result, instead, found no difference between experts and non experts for the inversion of the dog pictures at all. Recent studies into domain pacifistic show particular activation in the site of the FAA (fusion face area) from fem. scans, this is an area also implicated in propagandist when damaged, this area responds twice as strongly to faces as to other object classes, it also reflects the classic behavioral markers of face processing, being greater sensitivity to differences between individual upright faces then differences between inverted faces. In addition to the fem. scans, further studies in monkeys using the ultimate high resolution method found that 97% of visually responsive neurons in this region were strongly face selective.
Whereas the Expertise model suggests that children needed ten years of experience of facial exposure to develop the ‘hallmarks’ later used for recognition, the domain specificity theory suggests the face template has developed by one of two suggestions, through evolutionary processes by 4 stages, 1. An innate template with the basic structure of a face must be developed, secondly, the grouping of neurons in the FAA for good facial recognition, thirdly, activation of this template must rely on appropriate input during sensitive period, for example, a normal amount of facial exposure urine infancy in a standard upright position, explaining difficulty in adults recognizing inverted faces.
The second suggestion is more vague but in keeping with the necessity for biased exposure to faces in infancy, without any sort of innate face template, but from a different factor, hence the name ‘the infant experience plus other factor theory, although the other factor is not clear, it accounts for variability’s in recognition. McKeon et al argue that if the expertise theory were true and that face processing becomes special through continuous exposure to faces and that he skill is no different from other forms of visual ‘expertise’ practiced though learning, people have an ample opportunity to develop expertise with such other parts like hands or body shape, adults fail to do this and remain poor at identifying these parts compared to the face.
They state that it IS clearly not genetic expertise as if it were, then objects of expertise should be processed in the same way as faces, instead, they argue, there is significant evidence from the use of fem. scans that the adult visual systems contain specific systems tuned to recognize faces. The confounding evidence for the Expertise hurry lies in Diamond and Caress dog expert study in which has not been found to replicate the same results in any further research. Although the two theories are different in approach, they do share the ‘attention explanation’ and suggest that efficient exposure to faces in early infancy is necessary for the development of recognition in adult life, McKeon et al suggest this equally explains any weak expertise effects observed in the FAA for non face objects.