innovator

Add idea


Calendar

«    May 2019    »
MonTueWedThuFriSatSun
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
 

 

Advert

 

Payment

 

Advert

 

Authorization

Стартап

DateDate: 29-05-2019, 06:02


 Personal non-property and property rights for the invention belong to the author and protected under current legislation.
 The idea is that in a cot vstanovleni sensors that respond to a number of characteristics of the child aged till 3 years(or more). 

1. Cot is able to "pump up" of the child at a certain movement during sleep. Young parents will understand, when from the crib you have not to retreat, shaking it if necessary. Or when programming , the crib can swing in any desired period of time. 

2 . Automatic change diapers in the crib, allowing the baby lay without a diaper. 

3. Automatic extension reclining seats to beds , more comfortable maintenance of the child .

4. Automatic movement of the crib at home using pre-programmed home navigation. 

5. Climate system baby cot and baby monitors system. 

6. And of course, the relationship of the crib with a smartphone using the appropriate app.

7. The inner wall beds have the ability to pump air, to reduce injury to the child.
  The idea is open for investment and complete its foreclosure.
  +380505238948 


DateDate: 29-05-2019, 05:58

Experts have taught technology to draw with the help of several million videos.
A group of American developers introduced the Speech2Face neural network, capable of drawing a portrait of a man according to his voice. Reported arXiv.org.
Models are able to draw an image, based on gender, race, age. The team engaged in the development of the Massachusetts Institute of Technology, which included Tahyon O.
Scientists decided to use AVSpeech for learning neural network. It contains about a million short videos of over a hundred thousand different people. In this case, the video and audio tracks are separated.
The accuracy of the technology makers determined according to three demographic indicators. Comparison was made of gender, approximate age and race of people from the original clips and the "conclusion" made by the neural network.
Were also identified the shortcomings of the model. She is not always able to determine age with an accuracy of ten years, and best of all depicts Caucasians and Asians. It is believed that the latter problem is related to the uneven presence of different races in the sample for training.
According to the researchers, their plans did not include a thorough copying of a person’s appearance. They sought to accurately identify gender, age, ethnic group.
In the photo: the original image, restored and "drawn" by voice.