A powerful imaging algorithm that is able to see beyond what humans are capable of has been released as open source by the University of California, Los Angeles (UCLA), with uses for it in autonomous vehicles and security touted.
The UCLA engineering research group led by Prof Bahram Jalall had been working on the algorithm, called the Phase Stretch Transform (PST) algorithm, for some time, with it having some pretty impressive and far-reaching potential.
According to the university, the algorithm came following research into a technique called ‘photonic time stretch’, which has been used previously as a method of detecting cancerous cells in a blood sample using an ultrafast, light-sensitive camera.
From this, the research team took this technology and developed the PST algorithm, which has, in effect, expanded its possible uses to everything from autonomous vehicle camera software to high-end biosecurity.
In a practical sense, the algorithm looks at an object, targeting its outline, and then works inwards to find the fainter definitions on the object and make them ‘pop out’ to the camera.
Taking the example given by the university of the lamp, the algorithm is capable of looking past the noise of the light and see the actual bulb emitting it.
It can also be seen as a potential method of stargazing, as the obscuration of the night sky by the light reflecting off the moon can be filtered out, allowing the stars behind it to become more visible.
The algorithm is now available for download from Github and Matlab Exchange, with UCLA saying its release as open source will help improve it, as well as hoping to see it become a standard in image-processing applications in the future.
Camera lens image via Shutterstock
Buy your tickets now!