Title :
CNN computational abilities of large infrared proximity arrays
Author :
Koller, M. ; Cserey, G.
Author_Institution :
Fac. of Inf. Technol., Pazmany Peter Catholic Univ., Budapest, Hungary
Abstract :
In this paper we introduce a new hardware-software system, which applies some special modification on the original local connection attributes. The cells of this device has reflective type infrared distance sensors, each cell containing a LED and a phototransistor. The mapped surface can be not only the input of the computational system, but can form the local weights of the connections of adjacent cells. In this way, the locally adapted weightening is accomplished by the reflectance feature, which feature can modify the light of the neighboring cells. It has an important meaning: the physical environment take a part (and probably the control) of the processing. With this method, instead of using "distance image" as an input only, we can achieve more complex processing.
Keywords :
cellular neural nets; image processing; infrared detectors; light emitting diodes; phototransistors; sensor arrays; CNN; LED; cellular neural net; hardware software system; infrared distance sensors; infrared proximity array; phototransistor; reflectance feature; Cellular neural networks; Computational modeling; Infrared sensors; Light emitting diodes; Microcontrollers; Optical computing; Phototransistors; Sensor arrays; Sensor phenomena and characterization; Sensor systems;
Conference_Titel :
Cellular Nanoscale Networks and Their Applications (CNNA), 2010 12th International Workshop on
Conference_Location :
Berkeley, CA
Print_ISBN :
978-1-4244-6679-5
DOI :
10.1109/CNNA.2010.5430307