Abstract
In this work, we construct a model of an asymmetrically coupled network of semiconductor chaotic lasers in order to recognize noisy digital images of digits 0–9, derived from different samples in the digital image sets 0–9 found within the MNIST dataset. Here, the lasers network consists of eight asymmetrically coupled semiconductor lasers. The chaotic lasers network is driven by the external inputs, which encode one noise digital image to be recognized. The outputs of the chaotic lasers network driven by a total of 40 samples from the digital image sets 0–9 are utilized as ten sets of reference signals. The output of the chaotic lasers network induced by one noisy digital image is used as a test signal. By judging the maximum of the correlations of the test signal with the ten sets of reference signals, all noisy digital images 0–9 can be recognized well under different noises. Moreover, we further explore the recognition rate for each noisy digital image under different noises and a fixed injection strength. It is found that all noisy digital images can be recognized well under a certain low injection strength. The recognition-rates of all noisy digital images can further decrease to a certain extent under higher noise and a fixed the injection strength. The injection strength has little influence on the recognition rate of one noise digital image target with lower noise. The recognition rate under higher noise maintains a higher value (more than 0.9) when the injection strength is smaller than a certain value, but for the larger injection strength, the recognition rate exhibits further decrease. The modeled chaotic lasers network can play the role of photonic accelerators for the recognition of the noisy digital images.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.