US 12,474,770 B2
Eye tracking device, eye tracking method, and computer-readable medium
Thomas Debrunner, Zürich (CH); Pierre Giraud, Opfikon (CH); Chenghan Li, Zürich (CH); and Kynan Eng, Zürich (CH)
Assigned to INIVATION AG, Zürich (CH)
Appl. No. 18/004,152
Filed by INIVATION AG, Zürich (CH)
PCT Filed Jun. 30, 2021, PCT No. PCT/EP2021/067972
§ 371(c)(1), (2) Date Jan. 3, 2023,
PCT Pub. No. WO2022/003013, PCT Pub. Date Jan. 6, 2022.
Claims priority of application No. 20184020 (EP), filed on Jul. 3, 2020.
Prior Publication US 2023/0266818 A1, Aug. 24, 2023
Int. Cl. G06F 3/01 (2006.01); G06T 7/70 (2017.01); H04N 25/47 (2023.01)
CPC G06F 3/013 (2013.01) [G06T 7/70 (2017.01); H04N 25/47 (2023.01); G06T 2207/20084 (2013.01); G06T 2207/30201 (2013.01)] 13 Claims
OG exemplary drawing
 
1. An eye tracking device comprising:
an event-based optical sensor, which is configured to;
receive radiation reflected off an eye of a user and produce a signal stream of events, each event corresponding to detection of a temporal change in the received radiation at one or more pixels of said optical sensor; and
a controller, which is connected to said optical sensor and configured to:
a) receive the signal stream of events from said optical sensor,
b) convert a portion of said signal stream of events into a sparse tensor,
c) use said sparse tensor as an input for a first artificial neural network which is a recurrent neural network (RNN) having at least one memoized layer,
d) generate an inference frame utilizing said first artificial neural network based on at least said portion of said signal stream of events,
e) utilize said inference frame as input to a machine learning module and operate said machine learning module to obtain output data, and
f) extract from said output data information related to said eye of said user,
wherein only the values in the RNN that depend on a tensor element that is non-zero in said sparse tensor are updated.