From today you are also a rapper!Both the cat and the instrument copied with the camera are wrapped in AI "GIORGIO CAM" and the tension is improved.

The screen that tried Giorgio CAM

　"EXPERIMENTS WITH Google" is a showcase that introduces experimental applied examples to show the latest technologies such as artificial intelligence (AI) and expansion reality (AR).It has released a huge amount of content, and many of them can be tried on smartphones and PCs.In this article, we picked up interesting things from a wide variety of application examples.Through the experience report you actually played, we will explain the technology on the back side.I would be glad if the readers could learn the mechanism while trying them out with their own hands.You will surely notice the charm of the latest technology.

"GIORGIO CAM" that allows you to enjoy improvisation lap using a camera

　This time, we will cover "GIORGIO CAM", which can be a rapper if you have a smartphone.GIORGIO CAM is a web application that uses smartphones and laptop cameras.AI recognizes the object copied with the camera with AI, automatically generates wrap -like lyrics as a subject, and sings on a beat.Should it be called "lyrics" instead of "lyrics" because it's a wrap?It is very easy to use.As it works on a web browser, just access the Giorgio Cam site and tap "Launch Experiment".Once the loading is completed, press the "Let's go" button to play immediately.When the camera starts out and the camera starts, let's catch the nearby object on the screen.The image is judged by AI and improvisation of the rap on the theme.The effect of the screen also feels like a DJ floor lighting.Just taking a picture will allow anyone to play an improvised rap play according to the spot.It is purely fun, and there is also a good thing to experience the techniques such as AI's object identification and speech synthesis.It's exactly like Experiments with Google.

Even if it is recognized, the tension goes up on the wrap!

　Let's play immediately.It seems to work with a notebook PC and Android smartphone, but it was hard to swing the PC, so I made it a smartphone.However, with the "Pixel 3" used this time, it could only be shot with the inside camera, making it difficult to catch the subject.Moreover, since the image displayed on the screen does not flip left and right, it is difficult to fit well in the angle of view.In the sample video, it seems convenient because the outer camera can be used.I don't know if it depends on the model or the Android OS version, but I had a hard time getting used to it.Small things are hard to be recognized in the inside camera, so it is difficult to carry and shoot, so I put my smartphone on the stand.First of all, what was close.Once you start, you can play 5 times in a row.He will sing two answers at one time.The first watches were "Sleeve" (sleeves) and "Gesture" (gestures), and the second fountain pen was "Hand" (hand) and "Finger" (finger).The third largest glasses were interpreted as "VISION CARE" (visual acuity supplements) instead of "glasses" (glasses), and then answered "Thumb" (thumb).。It seems that he is not confident, and the displayed reliability is not as high as 80 %.When I showed the scissors fourth, it became "Tableware" (tableware) and "nail" (nails).Did you make a mistake as a hand with a tong?Finally, when I copied the miniature of the "Tower of the Sun", he sang "TOY" (toy) and "Creative Arts" (art work).Finally, it's safe to get the right answer.I'm not good at small things and are influenced by other prominent things on the screen.However, incorrect recognition is interesting because of misunderstanding, and applause seems to have been excited on the floor.If it's a big one, it's almost all questions. Next, I tried it with a little bigger.Then, except for the fourth of the five of the five, the answer was correct.The answers returned by the camera and the Giorgio CAM are as follows.1.Cup: Drinkware (drink tableware), Tableware (tableware) 2.Wine bottle: Bottle (bottle), Wine (wine) 3.Camera: Photographer (photographer), Digital Camera 4.Headphones: ARM (arm), Gesture 5.Laptop: Computer (computer), Laptop (notebook PC) The detailed differences between the instruments were changed to the irregular hair color and tried the instrument at hand, but the correct answer rate was high because it was large enough.The only mistake was the first tambourine, and I was dragged by "ARM" and "Sleeve" on the screen.The second synthesizer was "Musical Instrument" (instrument) and "Keyboard" (keyboard).By the way, when I showed the piano at another timing, I answered "piano", and it turned out that both were properly distinguished.However, the acoustic guitar, electric guitar, and electric bass from the third could not be distinguished as follows.1.Tambaline: ARM, Sleeve2.Synthesizer: Musical Instrument, Keyboard3.Acoustic guitar: GUITAR, String Instrument (stringed instrument) 4.Electric Guitar: MUSICIAN, Plucked String Instrument 5.Electric bass: Hand, String Instrument Accessory (Parts of stringed instruments) What is the distinction between cats and stuffed animals?Can creatures and stuffed animals be distinguished?I've tried.When he copied the cat, he answered "CAT" (cat) correctly, but also twisted "carnivore" (carnivorous animal) and "vertebrate" (spinal animal).It seems that it is surprisingly difficult to judge only a part like "LEG" (leg).I have answered "Comfort" (comfortable), and it seems like a cat like a cat, but what kind of learning did AI learn?Dumbo was famous worldwide, and while correctly answered "toy" and "Stuffed toy", he misunderstood "Hand" and "Watch" (watch).The Czech national character "Kurtech" and "Noraneko", which appeared in the popular "11 Pikinoneko" in Japan, had a delicate result.In addition to "Hand" and "ARM", it was displayed as "Joint" (joints) and "SHOULDER" (shoulder), which was significantly affected by the form.The Noraneko seems to have been stuck in a prominent part, such as "HEAD" (head), "Product" (product), and "EYE" (eyes).