Add idea


«    December 2018    »










DateDate: 17-12-2018, 06:01

 Personal non-property and property rights for the invention belong to the author and protected under current legislation. 
 In the author's ideas smart lapel-badge"Innovator" will be shown quite an ambitious project, which will help you to use memories of yesteryear when we collected and wore badges with the integration to date in relation to smart devices. The integration will help to extend the scope of wearable mobile devices, now in relation to the author's project in the form of a smart lapel-badges.
 Revealing the deeper the functionality of this smart device, you can designate its various use and functionality. First badge smart lapel-badge will be applicable to your core purpose with the help of the eponymous mobile app in Your gadget, the user can display on the touch screen smart lapel-badge various types and the most common lapel available icons, which will be located in the library of this mobile app. That is, the user chooses in the app badge and displays it via the usual pairing on the screen of your smart lapel-badge. Also on the smart lapel-badge will be able to withdraw any of such information, for example the brand of the clothes, which is wearing by the user. 
 Smart lapel-badge "Innovator" as an extra feature is built-in video camera. This will allow you to shoot photos and videos in real-time. For convenience, two or more interlocutors, will be able to functional applications the smart lapel-badge to evaluate the emotional state at the time of their dialogue, and display emoticons different appearance. For example, when the Jolly laughter during the dialogue smart lapel-badge will show the happy faces, thereby increasing a truly emotional atmosphere in this time and in this company.
 One of the functions is an automatic scanning barcode, which increasingly appear on all promotional materials. When the scan is completed, the processed information is displayed on the screen of Your gadget. Naturally, all existing apps will be available for use in this smart device that will be functional to approach it. For example, you can use the Skype app when to remove from clothing smart lapel-badge and put it in front of himself.
 The shape and size will be completely different. Charging will take place as a standard way and using the new wireless charging technology. In a separate equipment will be flashlight, microphone, and speaker.
 So here we have a brand new, innovative, wearable smart device that can be a fashionable mobile accessory will be equipped with a range of useful functions and will simplify the use of a mobile gadget, as the number of functions will go to him. But the main functionality will remain displaying lapel badges in active mode.
 This idea is open to investment, and full of redemption.

DateDate: 17-12-2018, 05:56

The generative adversary neural network learned how to create photos of high-resolution people with the smallest details - skin pores, wrinkles, moles, pigment spots, etc. However, when the same mechanism was tried on cats, it was not so convincing.
The company Nvidia, one of the most active players in the market of artificial intelligence systems, taught the algorithm to generate very similar to real photos of people. The images are distinguished by high resolution, a variety of faces and attention to detail. The research work is published on the Arxiv website.
The mechanism of generative competitive neural network was used to create images.
This approach means that one part of the algorithm, trained on real photos, tries to create images until another part starts to confuse them with real ones.
The results of this competition are very impressive: the neural network creates images of people of different ages, races, "removes" faces from different angles, realistically imposes shadows.
5871e8dac-5af5-466d-ac16-50262b62e7ec.jpeg (58 KB)
But most importantly, as the creators emphasize, she understands the elements of images. The face for her is not just a set of definitely colored pixels. The algorithm understands whether the photo has glasses, bald patches, beard, freckles, how high the forehead is, and how the cheekbones are outlined.
6ef5b76a2-6a3d-4610-9243-137773092974.jpeg (61 KB)
To demonstrate how picky and sensitive the neural network is, Nvidia released a video. The first part shows how several real photos are “mixed” with another control image. From it, prototypes adopt facial expressions, hair color and race characteristics. In the second, how several real faces are “averaged” into one image with a maximum of similarities.
Futurism notes that for Nvidia, this is not the first attempt to create a "face generator". One of the earliest developments was presented in the fall of 2017. However, that AI was confused in proportions, and the photos had a low resolution.
The article also contains photos of cats generated by the new algorithm. They are not so perfect - on many images the proportions of animals, to put it mildly, are not perfect.
729bc6b32-4d88-460a-804b-332efee9108f.jpeg (70 KB)
In early December, Nvidia introduced another impressive image processing technology using AI: the machine was taught to transform the real environment into objects of the virtual world. If in the next version the specialists will work on the details, then it will not be necessary to make three-dimensional models and spend time on rendering.