Live Updates: COVID-19 Cases
  • World 21,357,771
    World
    Confirmed: 21,357,771
    Active: 6,443,095
    Recovered: 14,151,295
    Death: 763,381
  • USA 5,476,266
    USA
    Confirmed: 5,476,266
    Active: 2,429,584
    Recovered: 2,875,147
    Death: 171,535
  • Brazil 3,278,895
    Brazil
    Confirmed: 3,278,895
    Active: 788,022
    Recovered: 2,384,302
    Death: 106,571
  • India 2,527,308
    India
    Confirmed: 2,527,308
    Active: 668,618
    Recovered: 1,809,542
    Death: 49,148
  • Russia 912,823
    Russia
    Confirmed: 912,823
    Active: 174,361
    Recovered: 722,964
    Death: 15,498
  • South Africa 579,140
    South Africa
    Confirmed: 579,140
    Active: 105,850
    Recovered: 461,734
    Death: 11,556
  • Peru 516,296
    Peru
    Confirmed: 516,296
    Active: 136,208
    Recovered: 354,232
    Death: 25,856
  • Mexico 511,369
    Mexico
    Confirmed: 511,369
    Active: 109,808
    Recovered: 345,653
    Death: 55,908
  • Chile 382,111
    Chile
    Confirmed: 382,111
    Active: 16,734
    Recovered: 355,037
    Death: 10,340
  • Spain 358,843
    Spain
    Confirmed: 358,843
    Active: 330,226
    Recovered: ?
    Death: 28,617
  • Iran 338,825
    Iran
    Confirmed: 338,825
    Active: 25,683
    Recovered: 293,811
    Death: 19,331
  • UK 316,367
    UK
    Confirmed: 316,367
    Active: 275,009
    Recovered: ?
    Death: 41,358
  • Saudi Arabia 295,902
    Saudi Arabia
    Confirmed: 295,902
    Active: 29,605
    Recovered: 262,959
    Death: 3,338
  • Pakistan 288,047
    Pakistan
    Confirmed: 288,047
    Active: 16,261
    Recovered: 265,624
    Death: 6,162
  • Bangladesh 271,881
    Bangladesh
    Confirmed: 271,881
    Active: 111,667
    Recovered: 156,623
    Death: 3,591
  • Italy 252,809
    Italy
    Confirmed: 252,809
    Active: 14,249
    Recovered: 203,326
    Death: 35,234
  • Turkey 246,861
    Turkey
    Confirmed: 246,861
    Active: 11,947
    Recovered: 228,980
    Death: 5,934
  • Germany 223,774
    Germany
    Confirmed: 223,774
    Active: 11,935
    Recovered: 202,550
    Death: 9,289
  • France 212,211
    France
    Confirmed: 212,211
    Active: 97,957
    Recovered: 83,848
    Death: 30,406
  • Canada 121,652
    Canada
    Confirmed: 121,652
    Active: 4,690
    Recovered: 107,942
    Death: 9,020
  • China 84,808
    China
    Confirmed: 84,808
    Active: 655
    Recovered: 79,519
    Death: 4,634
  • Netherlands 61,840
    Netherlands
    Confirmed: 61,840
    Active: 55,673
    Recovered: ?
    Death: 6,167
  • Australia 23,035
    Australia
    Confirmed: 23,035
    Active: 9,301
    Recovered: 13,355
    Death: 379
  • S. Korea 15,039
    S. Korea
    Confirmed: 15,039
    Active: 833
    Recovered: 13,901
    Death: 305
  • New Zealand 1,609
    New Zealand
    Confirmed: 1,609
    Active: 56
    Recovered: 1,531
    Death: 22

New Google algorithm makes robots better equipped at recognizing transparent objects

Author at TechGenyz Google
Google Transparent Objects

ClearGrasp, a new learning algorithm has been developed by researchers at Google Synthesis AI, and Columbia University, which will help robots interact with transparent objects. The algorithm uses RGB-D images to recreate the 3D spatial information of the object in question.

Robots make use of the RGB-D cameras to paint an accurate 3D picture of the environment that it is in. However, there certainly are limitations to the surroundings created by such cameras; for example, it does not work effectively for transparent objects such as glass.

To recreate 3D spatial information for transparent objects proved itself to be a herculean task for the researchers. There was very little data available meant for transparent surfaces, and most of the data blatantly ignored the transparent surfaces. To overcome this issue, the researchers created a large-scale transparent object data set, containing 50,000 realistic renderings of various object surfaces.

The ClearGrasp algorithm uses 3 neural networks to correctly identify transparent objects. One of the networks estimates the surface normal vector, one of them calculates the edge of the occlusion and the other one calculates the transparency of the object. The mask of the object is used to exclude pixels of non-transparent objects in order to fill the correct depth.

The global optimization module can predict the normal vectors of other surfaces from surfaces of known depth to reconstruct the shape of the object and to differentiate between two objects.

However, the algorithm could not correctly detect the normal vectors of other basic surfaces, due to the limitation of the synthetic data set. To tackle this problem, the researchers came up with the Matterport3D and ScanNet data set.

However, all of these mishaps aside, ClearGrasp is the only algorithm available that can reconstruct the depth of transparent objects, increasing the success rate of grasping transparent objects by the robotic arm from 12% to 74%.

Career

Subscribe