Traditional Culture Encyclopedia - Photography and portraiture - Can the artificial intelligence seen at the Google conference judge whether the photos are good or not?
Can the artificial intelligence seen at the Google conference judge whether the photos are good or not?
However, after the appearance of AlphaGO, I believe many people also understand the potential of "artificial intelligence". They should undertake more tasks and create more value for mankind. The artificial intelligence Watson introduced by IBM has been able to complete the tasks of music creation and text writing. With creativity, it is more like human beings. Having said that, some artificial intelligence products are creative, but it is still difficult for them to judge the quality of their works. To make artificial intelligence have the same judgment as people, we must let them learn more.
Recently, Google released a system called Neuron Image Assessment System (NIMA). This system can judge the "good" and "bad" photos, and the scores it gives are very close to those evaluated by people.
According to Engadget's report, Google's NIMA system has the ability of deep learning. It can learn from others' comments on photos, and then integrate these data to form a complete evaluation system.
This evaluation system can make analysis according to the combination of composition, color and object position, and then make judgments according to the collected data, so as to grade your photos.
Google also found some professional photo commentators in related tests. They compared the results calculated by NIMA with those calculated by commentators and found that there was not much difference between them. It can be seen that NIMA system can still "learn aesthetics" according to the data given by human beings, and the comments should be close to the "public taste", or it can be realized by piling up a lot of data. Although Google did not disclose the judgment method of this system, they mentioned in the research report that NIMA can actually reflect the personalized style. As long as the input data and the photo style of comments are unified, NIMA will form a specific style in this learning process. If it is an artificial assistant implanted in your mobile phone and faces your photos for a long time, then its final aesthetic style will be consistent with yours.
In fact, this system has certain significance for future creation. NIMA can judge a work as "good" or "bad" and can adapt to your style through continuous learning. In this way, it can not only comment on your photos, but also help you when shooting. Integrating photo criticism skill NIMA with photo assistance, since it learns photo criticism, let it learn photography skills. When the user opens the viewfinder, artificial intelligence can identify the current picture and composition, so as to select the current suitable scheme more accurately. This is also a good solution for some users who don't know how to take pictures at all. At this time, the existence of artificial intelligence is to reduce the learning accumulation cost of users, and its assistance can make people who have not learned this technology get started quickly and take "not bad" photos. Although this will reduce people's thinking and personal style. But after all, this corresponds to users who have never studied, and it is not a bad thing to take a good-looking photo.
In addition, Engadget also mentioned in the report that NIMA can be used in some post-photo applications, which will bring more suggestions for users to retouch the photos later. After learning, it can also adjust the color and brightness contrast of photos to suit your taste, instead of just using mathematical models to adjust the curve, which is also a very convenient thing. Generally speaking, the appearance of NIMA has brought a new direction to film production. Or, when we search for pictures on Google in the future, we can directly see the scores detected by NIMA, or your mobile phone will jump out and think that your photos are too bad. This should not be far away.
- Previous article:Kangding travel guide must-see attractions
- Next article:Swivel chair photography
- Related articles
- Is it allowed to take a SLR camera to watch the ball at Bird's Nest National Stadium?
- What gift should I give my friend's shop for the tenth anniversary celebration?
- Tianqiaogou Forest Park
- Swiss photographers filmed Tibet.
- What is an objective lens?
- High score! Give my son a name
- There seems to be a mildew spot in the polaroid lens. What should I do?
- How to write the artistic practice of comprehensive quality evaluation?
- What about Hengyang Hong Xin Property Management Company?
- What are the majors of Xi 'an University of Arts and Sciences?