Apparatus and methods for detecting a property from electromagnetic radiation sensor data

1. An apparatus comprising:

a sensor module comprising:

an electromagnetic radiation sensor configured to generate electromagnetic radiation sensor data;

a coded mask configured to modulate electromagnetic radiation incident to the electromagnetic radiation sensor and from which the electromagnetic radiation sensor data is generated; and

at least one processor configured to:

obtain the electromagnetic radiation sensor data from the electromagnetic radiation sensor,

detect a property from the electromagnetic radiation sensor data using an artificial neural network, and

output information related to the detected property via an output,

wherein the sensor module further comprises a filter configured to block electromagnetic radiation outside a pre-defined frequency range.

2. The apparatus according to claim 1 , wherein:

the electromagnetic radiation sensor data corresponds to raw electromagnetic radiation sensor data, and

the at least one processor is configured to detect the property using the artificial neural network without computing an image representation of the electromagnetic radiation sensor data.

3. The apparatus according to claim 1 , wherein the sensor module is a lensless sensor module.

4. The apparatus according to claim 1 , wherein the sensor module and the at least one processor are arranged according to at least one of:

the sensor module and the at least one processor are implemented on a single chip,

the sensor module and the at least one processor are implemented within or on a same semiconductor substrate,

the sensor module and the at least one processor are implemented within a same semiconductor package, or

the sensor module and the at least one processor are implemented using a chip stack.

5. An apparatus comprising:

a sensor module comprising:

an electromagnetic radiation sensor configured to generate electromagnetic radiation sensor data;

a coded mask configured to modulate electromagnetic radiation incident to the electromagnetic radiation sensor and from which the electromagnetic radiation sensor data is generated; and

at least one processor configured to:

obtain the electromagnetic radiation sensor data from the electromagnetic radiation sensor,

detect a property from the electromagnetic radiation sensor data using an artificial neural network, and

output information related to the detected property via an output,

wherein the coded mask is an aperture-coded mask, and

wherein the aperture-coded mask is one of a Modified Uniformly Redundant Array, a Hexagonal Uniformly Redundant Array, a pinhole, an aperture-coded mask with a pseudo random pattern, an aperture-coded mask with a random pattern, an aperture-coded mask with a chip-specific or chip-individual pattern, an aperture-coded mask having a near-flat Fourier spectrum, or an aperture-coded optical mask.

6. The apparatus according to claim 1 , wherein the coded mask is a grate-coded mask.

7. The apparatus according to claim 1 , wherein:

a distance between the electromagnetic radiation sensor and the coded mask is less than 5 mm,

the electromagnetic radiation incident to the electromagnetic radiation sensor is at least one of ultraviolet light, visible light, or infrared light, and

the property is an optical property of at least one object in a vicinity of the apparatus, wherein the electromagnetic radiation incident to the electromagnetic radiation sensor is influenced by the at least one object.

8. The apparatus according to claim 1 , wherein the electromagnetic radiation sensor is configured to sense at least one of visible light, infrared light, or ultraviolet light.

9. The apparatus according to claim 1 , wherein the filter is an optical filter configured to block light outside at least one of a visible light frequency range, an infrared light frequency range, or an ultraviolet light frequency range.

10. The apparatus according to claim 1 , wherein the sensor module further comprises a tunable color filter configured to block light outside an adjustable frequency range.

11. The apparatus according to claim 1 , wherein the artificial neural network is implemented using resistive random-access memory elements to configure its function.

12. The apparatus according to claim 1 , wherein:

the at least one processor is configured to process the detected property to generate a control signal to control an external entity, and

the at least one processor is configured to emulate the artificial neural network.

13. The apparatus according to claim 1 , wherein:

the at least one processor is configured to classify an optical property of at least one object influencing the electromagnetic radiation incident to the electromagnetic radiation sensor using the artificial neural network, and

the at least one processor is configured to perform at least one of:

facial recognition or facial interpretation based on the electromagnetic radiation sensor data using the artificial neural network,

gesture detection or gesture interpretation based on the electromagnetic radiation sensor data using the artificial neural network,

motion detection or motion interpretation based on the electromagnetic radiation sensor data using the artificial neural network, or

determination of an image representation of the electromagnetic radiation incident to the electromagnetic radiation sensor based on the electromagnetic radiation sensor data using the artificial neural network.

14. The apparatus according to claim 1 , wherein the artificial neural network comprises at least two hidden layers.

15. The apparatus according to claim 1 , wherein the at least one processor is configured to at least partially train the artificial neural network based on the coded mask, or

wherein the artificial neural network corresponds to a pre-trained artificial neural network.

16. A method comprising:

sensing modulated electromagnetic radiation modulated by a coded mask to generate electromagnetic radiation sensor data;

blocking electromagnetic radiation outside a pre-defined frequency range to obtain filtered electromagnetic radiation, wherein the electromagnetic radiation outside the pre-defined frequency range is filtered from the modulated electromagnetic radiation or from pre-modulated electromagnetic radiation;

detecting a property from the electromagnetic radiation sensor data using an artificial neural network; and

outputting information related to the detected property.

17. A method for manufacturing an apparatus, the method comprising:

forming a sensor module by:

forming an electromagnetic radiation sensor configured to provide the electromagnetic radiation sensor data;

forming a coded mask configured to modulate the electromagnetic radiation incident to the electromagnetic radiation sensor; and

forming at least one processor comprising an artificial neural network, wherein the at least one processor is configured to:

obtain the electromagnetic radiation sensor data from the electromagnetic radiation sensor,

detect a property from the electromagnetic radiation sensor data using the artificial neural network, and

output information related to the detected property,

wherein forming the at least one processor further comprises at least partially training the artificial neural network using the at least one processor based on the coded mask, or

wherein forming of the at least one processor further comprises loading information related to a pre-trained artificial neural network into the at least one processor to obtain the artificial neural network.

18. The method according to claim 17 , wherein forming of the sensor module comprises:

forming the electromagnetic radiation sensor within a semiconductor substrate of a semiconductor device comprising the apparatus;

forming the coded mask on or within the semiconductor substrate; and

forming a cavity between the electromagnetic radiation sensor and the coded mask, wherein a vertical distance between the electromagnetic radiation sensor and the coded mask within the cavity is less than 1 mm.

19. The apparatus according to claim 1 , wherein:

the coded mask is an aperture-coded mask, and

the aperture-coded mask is one of a Modified Uniformly Redundant Array, a Hexagonal Uniformly Redundant Array, a pinhole, an aperture-coded mask with a pseudo random pattern, an aperture-coded mask with a random pattern, an aperture-coded mask with a chip-specific or chip-individual pattern, an aperture-coded mask having a near-flat Fourier spectrum, or an aperture-coded optical mask.

20. The apparatus according to claim 5 , wherein the sensor module further comprises a filter configured to block electromagnetic radiation outside a pre-defined frequency range.