On the evening of August 24th, Beijing time, Google is developing a new technology that uses neural networks to simulate the working methods of the human brain to compress images. This technology can greatly reduce the size of the image file while ensuring the quality of the image.
This new technology is of great significance. For consumers, the smaller file size means more space on their phones, tablets and PCs. For technology companies such as Google that offer photo storage services, smaller photos can reduce server load and power consumption, as well as speed up delivery.
According to Google researchers, this new technology is more efficient than traditional JPEG image compression technology. It is reported that the researchers used 6 million randomly selected photos to train artificial intelligence systems to learn how to compress images.
The researchers broke down the photos into small blocks of 32x32 pixels, and then selected 100 of the smallest compression-efficient images for the system to learn. Google's intention is to use the most difficult pieces to train, the compression of other parts of the image will become very simple.
Next, artificial intelligence will predict what the compressed image will look like and generate the result. In addition to the compressed file size smaller than the JPEG format, the neural network can also do the best compression for different parts of the same picture.
Of course, the system is not yet mature. When compressing a picture, sometimes the resulting image does not look perfect with the naked eye. Therefore, this technology may not be popular in the short term, but the progress made by Google is still impressive.
Disposable Oil Vaporizer,Ceramic Oil Vaporizer,Cbd Juice Vaporizer,Disposable Cbd Juice Oil Vaporizer
ALD GROUP LIMITED , https://www.aldvapor.com