|Photo courtesy of Blue River Technology
Blue River Technology, a young startup out of Stanford University, kills weeds using Computer Vision and machine learning. In the process, farmers maximize their yield and cut back on the use of herbicides.
Currently used for lettuce crops, the company’s technology learned how to recognize the plant by analyzing close to 1,000,000 photos of lettuce. When a camera relays images of weeds growing among the lettuce plants, the software instructs a mechanical knife to root them out. As a backup, the software can send a signal to a sprayer that douses the weeds with herbicide.
In a few years’ time, with technology like this, could you see yourself growing the garden of your dreams in your backyard, freed from pulling weeds?
This blog is sponsored by ImageGraphicsVideo, a company offering ComputerVision Software Development Services.
While doctors have a varying success rate of detecting and diagnosing breast cancer, researchers at Stanford University have created a computer system known as the C-Path (Computational Pathologist), which examines tissue samples and is capable of diagnosing breast cancer as well as, if not better than, humans. Additionally, it is able to provide a likely prognosis.
Currently, doctors have to individually examine tissue samples of tumors under microscopes to determine if they’re affected. The C-Path cuts out the middle man, and in addition to serving as inspector, also has the ability to learn as it goes. In fact, its initial programming was based on being fed preexisting samples with known prognoses. This comparison of what the machine does versus the knowledge it was provided with allowed it to “learn” and adapt.
It is the hope of scientists that the C-Path’s abilities can be improved over time so that it has the ability to not only predict the chances of a patient’s survival, but also offer information as to which treatment would be most effective for a particular type of cancer.
Of course, in some situations there is no replacement for the care and exactness of a human inspection, but it is interesting to think what kind of strides a machine like this might make in the medical community.
Anyone who has typed an address into Google Street View has probably had the experience of seeing that the result shows the street, but not always the exact building at a specified address. This is because computer-vision software hasn’t been advanced enough to zero in on numbers. This is namely due to the fact addresses are displayed on buildings in a variety of colors, sizes and fonts, making it difficult for computers to pinpoint, recognize and extract information from them.
Photo courtesy of Netzer et. al.
However, researchers at Stanford University have teamed up with Google to improve the technology by creating an algorithm that’s able to more accurately identify street numbers. In “Reading Digits in Natural Images with Unsupervised Feature Learning,” the teams explain how they trained the system to recognize these numbers using computer-vision algorithms combined with technology that recognizes patterns and learns to adapt to and implement them.
It is the hope of Google that this kind of information could lead to a better Street View system, in addition to more accurate maps and navigation services.