![<?echo $_SERVER['SERVER_NAME'];?>](/template/twentyseventeen/skin/images/header.jpg)
 |
A review article entitled "The Tricky Problem Of Making Smart Fridges Smart" published by MIT "Technology Venture" magazine on Monday said that although smart refrigerator technology has not achieved a breakthrough, a group of South Korean scientists However, he hopes to improve the intelligence of this product through new methods and thus change this situation.
The following is the full text of the article:
Over the past 10 years, the linked refrigerator dream of the 1990s has always been laughed at. Even the most advanced smart refrigerators today are just adding little tricks to traditional refrigerators, such as barcode scanners or touch screens. It is no wonder that smart refrigerators cannot attract public interest.
However, Thomas Sandholm and his colleagues at the Korea Advanced Institute of Technology hope to change this situation. They have developed a system called CloudFridge that hopes to improve various deficiencies in smart refrigerators and give them a rightful place in the consumer technology market.
When they first conceived this concept in the 1990s, they hoped to use smart refrigerators to completely change the way we interact with food. The dream at the time was to use these devices to monitor the various foods that we put into and out of the refrigerator, and then warn when the food was reduced, or even purchase substitutes directly.
But this dream has never been realized. Sandholm said the main problem is that smart refrigerators require users to enter too much information. For example, many smart refrigerators require the user to scan the bar code one by one while putting food into and out of the refrigerator.
Of course, the failure does not stop there. Although some people have developed RFID (Radio Frequency Identification) scanners to track food usage, in fact, there is almost no food that uses RFID tags. There are some systems that can determine the frequency of use through the number of refrigerator door switches. This product also needs to be equipped with a distance sensor to detect if something is being taken out or put in the refrigerator. However, this system also failed because it could not determine the specific food or item type.
Although the history of smart refrigerators is not short, it has not yet been applied. This fully reflects the embarrassment of technology in terms of convenience and practicality.
Sandholm and his colleagues believe that they can change this situation through a prototype system that can learn and improve the user experience. "We hope to promote innovation in this area by evaluating real user perceptions."
In order to achieve this goal, they have developed a new refrigerator with three major advantages. First of all, this refrigerator incorporates an object recognition system that allows users to determine what to put in or take out of the refrigerator without the user's involvement.
The key to this function is to get Google's help. The refrigerator developed by the Sandholm team has a built-in webcam. Whenever something is in or out of the refrigerator, the camera sends the image to Google's image search service.
The second advantage is that the sensor measures the position of the object in the refrigerator.
The last advantage is the highlighting of objects in the refrigerator through a set of red or green LEDs that draw the user’s attention. These objects may need to be replaced in time, or they may be about to expire. The result of this prototype product is expected to far exceed the previous smart refrigerator.
The Sandholm team has developed several applications for use with this prototype product. One of the apps, called CloudFridge Take Out, can easily find items in the refrigerator through voice commands and use LEDs to highlight the object. The Sandholm team stated that the app can also highlight all the items in a particular menu to ensure that the user gets all the appropriate ingredients and ingredients.
"We also plan to add social features to share the contents of the fridge with friends or compare tastes so that we can use social channels and collective wisdom to get recommendations," they said.
The biggest problem is obviously object recognition. Even with the Google Image Search service, it still takes up to 5 seconds to analyze a specific picture and return results. Even so, the accuracy of the system still cannot reach 100%, and it is still difficult to identify the situation.
Sandholm's team has already encountered this situation in the test, which also led to the limited availability of the new system.
Object recognition technology will gradually improve, and Sandholm and his team have clearly taken the right step. However, the real smart refrigerator may not be able to enter millions of households in the short term. (books)
Check Valve,Swing Check Valve,Water Check Valve,Stainless Steel Check Valve
Wenzhou Dico Valve Technology Co.,Ltd. , https://www.dicovalves.com