Monitoring shopping activities using weight data in a store

1. A computer-implemented method for monitoring shopping activities, comprising:

receiving, by a computer system from a plurality of sensors including one or more image sensors, one or more pressure sensors, one or more capacitive sensors, and one or more weight sensors, data collected by the one or more image sensors, data collected by the one or more pressure sensors, data collected by the one or more capacitive sensors, and data collected by the one or more weight sensors;

identifying, by the computer system based at least on the data collected by the one or more image sensors, one or more product items removed by a person from the store with a first confidence, wherein the identifying the one or more product items comprises:

recognizing, by the computer system using a first trained machine-learning model, through an analysis of 3D motions of the person based on the data collected by the one or more image sensors, a motion of body joints of the person reaching for the one or more product items,

extracting, by the computer system from the data collected from the one or more image sensors, the data collected from the one or more pressure sensors, and the data collected from one or more capacitive sensors, a plurality of features associated with each of the one or more product items, wherein the features comprise a conductivity, a color, a position, and a surface force image, the conductivity is used to determine a type of material that makes up at least a part of the one or more product items, and the surface force image is used to determine a footprint of the one or more product items that includes a shape of a bottom surface of the one or more product items, and

determining, by the computer system using a second machine-learning model, an identifier associated with each of the one or more product items based on the extracted features including the conductivity, the color, the position, and the surface force image, wherein the second machine-learning model is trained based on training data collected from the one or more image sensors, the one or more pressure sensors, and the one or more capacitive sensors;

calculating, by the computer system, an expected total weight of the one or more product items based on information associated with the one or more product items stored by the computer system;

determining, by the computer system based on the data collected by the one or more weight sensors, an actual total weight of the one or more product items removed by the person; and

in response to the actual total weight being consistent with the expected total weight, verifying, by the computer system, that the one or more product items are removed by the person with a second confidence higher than the first confidence.

2. The method of claim 1 , wherein identifying one or more product items removed by a person from the store comprises:

extracting, by the computer system, form the data received from the one or more image sensors, one or more features associated with each of the one or more product items; and

determining, by the computer system using a machine-learning model, an identifier associated with each of the one or more product items based on one or more of the extracted features.

3. The method of claim 1 , wherein the calculating an expected total weight of the one or more product items based on information associated with the one or more product items stored by the computer system comprises:

storing, by the computer system in a data store associated with the computer system, weight information associated with a plurality of product items in the store;

retrieving, by the computer system from the data store, the stored weight information associated with the determined one or more product items removed by the person from the store; and

calculating the expected total weight based on the retrieved weight information.

4. The method of claim 1 , wherein the calculating an expected total weight of the one or more product items based on information associated with the one or more product items stored by the computer system comprises:

receiving, by the computer system from one or more sensors associated with one or more product-storage structures in the store, weight information associated with the one or more product items removed from the store by the person;

storing, by the computer system, the received weight information; and

calculating the expected total weight based on the stored weight information.

5. The method of claim 1 , wherein the determining an actual total weight based on data collected by the one or more weight sensors further comprises:

receiving, by the computer system from the one or more weight sensors, a first weight associated with the person at a first point of time;

receiving, by the computer system from the one or more weight sensors, a second weight associated with the person at a second point of time; and

calculating, by the computer system, the actual total weight based on a difference between the first weight and the second weight.

6. The method of claim 1 , wherein the identifying one or more product items removed by a person from the store comprises:

determining, by the computer system, one or more interactions between the person and the one or more product items based on the data collected by the one or more image sensors; and

determining, by the computer system, that the one or more product items are removed from the store by the person.

7. The method of claim 6 , wherein the determining one or more interactions between the person and the one or more product items comprises, for each of the one or more product items:

determining, based on the data received from the one or more image sensors, that the person is in proximity to the item in a period of time;

detecting one or more movements of a hand of the person with respect to the item; and

determining one or more interactions based on the one or more detected movements.

8. The method of claim 1 , wherein one or more of the one or more weight sensors are affixed to one or more structures in proximity to an entrance or an exit of the store.

9. A system for monitoring shopping activities comprising a computer system and a plurality of sensors, the computer system comprising one or more processors and one or more non-transitory computer-readable storage media storing instructions executable by the one or more processors to cause the system to perform operations comprising:

receiving, from a plurality of sensors including one or more image sensors, one or more pressure sensors, one or more capacitive sensors, and one or more weight sensors, data collected by the one or more image sensors data collected by the one or more pressure sensors, data collected by the one or more capacitive sensors, and data collected by the one or more weight sensors;

identifying, based at least on the data collected by the one or more image sensors, one or more product items removed by a person from the store with a first confidence, wherein the identifying the one or more product items comprises:

recognizing, using a first trained machine-learning model, through an analysis of 3D motions of the person based on the data collected by the one or more image sensors, a motion of body joints of the person reaching for the one or more product items,

extracting from the data collected from the one or more image sensors, the data collected from the one or more pressure sensors, and the data collected from one or more capacitive sensors, a plurality of features associated with each of the one or more product items, wherein the features comprise a conductivity, a color, a position, and a surface force image, the conductivity is used to determine a type of material that makes up at least a part of the one or more product items, and the surface force image is used to determine a footprint of the one or more product items that includes a shape of a bottom surface of the one or more product items, and

determining, using a second machine-learning model, an identifier associated with each of the one or more product items based on the extracted features including the conductivity, the color, the position, and the surface force image, wherein the second machine-learning model is trained based on training data collected from the one or more image sensors, the one or more pressure sensors, and the one or more capacitive sensors;

calculating an expected total weight of the one or more product items based on information associated with the one or more product items stored by the computer system;

determining, based on the data collected by the one or more weight sensors, an actual total weight of the one or more product items removed by the person; and

in response to the actual total weight being consistent with the expected total weight, verifying, by the computer system, that the one or more product items are removed by the person with a second confidence higher than the first confidence.

10. The system of claim 9 , wherein the identifying one or more product items removed by a person from the store comprises:

extracting from the data received from the one or more image sensors, one or more features associated with each of the one or more product items; and

determining, using a machine-learning model, an identifier associated with each of the one or more product items based on one or more of the extracted features.

11. The system of claim 9 , wherein the calculating an expected total weight of the one or more product items based on information associated with the one or more product items stored by the computer system comprises:

storing, in a data store associated with the computer system, weight information associated with a plurality of product items in the store;

retrieving, from the data store, the stored weight information associated with the determined one or more product items removed by the person from the store; and

calculating the expected total weight based on the retrieved weight information.

12. The system of claim 9 , wherein the calculating an expected total weight of the one or more product items based on information associated with the one or more product items stored by the computer system comprises:

receiving, from one or more sensors associated with one or more product-storage structures in the store, weight information associated with the one or more product items removed from the store by the person;

storing the received weight information; and

calculating the expected total weight based on the stored weight information.

13. The system of claim 9 , wherein the identifying one or more product items removed by a person from the store comprises:

determining one or more interactions between the person and the one or more product items based on the data collected by the one or more image sensors; and

determining that the one or more product items are removed from the store by the person.

14. The system of claim 13 , wherein the determining one or more interactions between the person and the one or more product items further comprise, for each of the one or more product items:

determining, based on the data received from the one or more image sensors, that the person is in proximity to the item in a period of time;

detecting one or more movements of a hand of the person with respect to the item; and

determining one or more interactions based on the one or more detected movements.

15. The system of claim 9 , wherein one or more of the one or more weight sensors are affixed to one or more structures in proximity to an entrance or an exit of the store.

16. A non-transitory computer-readable storage medium for monitoring shopping activities, configured with instructions executable by one or more processors to cause the one or more processors to perform operations comprising:

receiving, from a plurality of sensors including one or more image sensors, one or more pressure sensors, one or more capacitive sensors, and one or more weight sensors, data collected by the one or more image sensors, data collected by the one or more pressure sensors, data collected by the one or more capacitive sensors, and data collected by the one or more weight sensors;

identifying, based at least on the data collected by the one or more image sensors, one or more product items removed by a person from the store with a first confidence, wherein the identifying the one or more product items comprises:

recognizing, using a first trained machine-learning model, through an analysis of 3D motions of the person based at least on the data collected by the one or more image sensors, a motion of body joints of the person reaching for the one or more product items;

extracting from the data collected from the one or more image sensors, the data collected from the one or more pressure sensors, and the data collected from one or more capacitive sensors, a plurality of features associated with each of the one or more product items, wherein the features comprise a conductivity, a color, a position, and a surface force image, the conductivity is used to determine a type of material that makes up at least a part of the one or more product items, and the surface force image is used to determine a footprint of the one or more product items that includes a shape of a bottom surface of the one or more product items, and

determining, using a second machine-learning model, an identifier associated with each of the one or more product items based on the extracted features including the conductivity, the color, the position, and the surface force image, wherein the second machine-learning model is trained based on training data collected from the one or more image sensors, the one or more pressure sensors, and the one or more capacitive sensors;

calculating an expected total weight of the one or more product items based on information associated with the one or more product items stored by the computer system;

determining, based on the data collected by the one or more weight sensors, an actual total weight of the one or more product items removed by the person; and

in response to the actual total weight being consistent with the expected total weight, verifying, by the computer system, that the one or more product items are removed by the person with a second confidence higher than the first confidence.

17. The non-transitory computer-readable storage medium of claim 16 , wherein the identifying one or more product items removed by a person from the store comprises:

extracting from the data received from the one or more image sensors, one or more features associated with each of the one or more product items; and

determining, using a machine-learning model, an identifier associated with each of the one or more product items based on one or more of the extracted features.

18. The non-transitory computer-readable storage medium of claim 16 , wherein the calculating an expected total weight of the one or more product items based on information associated with the one or more product items stored by the computer system comprises:

storing, in a data store associated with the computer system, weight information associated with a plurality of product items in the store;

retrieving, from the data store, the stored weight information associated with the determined one or more product items removed by the person from the store; and

calculating the expected total weight based on the retrieved weight information.

19. The non-transitory computer-readable storage medium of claim 16 , wherein the calculating an expected total weight of the one or more product items based on information associated with the one or more product items stored by the computer system comprises:

receiving, from one or more sensors associated with one or more product-storage structures in the store, weight information associated with the one or more product items removed from the store by the person;

storing the received weight information; and

calculating the expected total weight based on the stored weight information.

20. The non-transitory computer-readable storage medium of claim 16 , wherein one or more of the one or more weight sensors are affixed to one or more structures in proximity to an entrance or an exit of the store.