Action detection in a video based on image analysis

1. A system comprising:

at least one processor and a computer readable storage medium having program instructions embodied therewith, the program instructions executable by the at least one processor to cause the at least one processor to perform operations comprising:

obtaining a plurality of images associated with an activity between a first person and a second person;

performing image analysis on the plurality of images to identify one or more body parts of the first person and one or more body parts of the second person;

using image analysis, tracking the one or more body parts of the first person and the one or more body parts of the second person over one or more images of the plurality of images;

based on the image analysis, detecting a physical contact between a given body part of the first person with a given body part of the second person in the one or more images of the plurality of images;

based on the image analysis, determining one or more positions of the given body part of the first person and the given body part of the second person at the physical contact in the one or more images of the plurality of images;

using the one or more positions, computing relative velocities of the given body part of the first person and the given body part of the second person at the physical contact;

based on the relative velocities and the one or more positions of the given body part of the first person and the given body part of the second person in the one or more images of the plurality of images, determining one or more characteristics of the physical contact; and

using a cognitive model, classifying the physical contact as a permitted interaction or an unpermitted interaction based at least on the one or more characteristics.

2. The system of claim 1 , wherein the plurality of images is obtained from a video.

3. The system of claim 1 , wherein the first object is a head of a first person, and wherein the second object is a torso of a second person.

4. The system of claim 1 , wherein the at least one processor further performs operations comprising determining the one or more characteristics associated with the physical contact, and wherein at least one of the characteristics is one of a movement direction, a source body part, a target body part, or relative angular direction between the given body parts of the first person and the second person making contact.

5. The system of claim 1 , wherein the at least one processor further performs operations comprising determining a match between the physical contact and one or more predefined interactions, and wherein one or more of the predefined interactions are unpermitted interactions.

6. The system of claim 1 , wherein the at least one processor further performs operations comprising augmenting the classifying of the physical contact as a permitted interaction or an unpermitted interaction based at least on sensor data.

7. A computer program product comprising a non-transitory_computer readable storage medium having program instructions embodied therewith, the program instructions executable by at least one processor to cause the at least one processor to perform operations comprising:

obtaining a plurality of images associated with an activity between a first person and a second person;

performing image analysis on the plurality of images to identify one or more body parts of the first person and one or more body parts of the second person;

using image analysis, tracking the one or more body parts of the first person and the one or more body parts of the second person over one or more images of the plurality of images;

based on the image analysis, detecting a physical contact between a given body part of the first person with a given body part of the second person in the one or more images of the plurality of images;

based on the image analysis, determining one or more positions of the given body part of the first person and the given body part of the second person at the physical contact in the one or more images of the plurality of images;

using the one or more positions, computing relative velocities of the given body part of the first person and the given body part of the second person at the physical contact;

based on the relative velocities and the one or more positions of the given body part of the first person and the given body part of the second person in the one or more images of the plurality of images, determining one or more characteristics of the physical contact; and

using a cognitive model, classifying the physical contact as a permitted interaction or an unpermitted interaction based at least on the one or more characteristics.

8. The computer program product of claim 7 , wherein the plurality of images is obtained from a video.

9. The computer program product of claim 7 , wherein the at least one processor further performs operations comprising determining the one or more characteristics associated with the physical contact, and wherein at least one of the characteristics is one of a movement direction, a source body part, a target body part, or a relative angular direction between the given body parts of the first person and the second person making contact.

10. The computer program product of claim 7 , wherein the at least one processor further performs operations comprising determining a match between the physical contact and one or more predefined interactions, and wherein one or more of the predefined interactions are unpermitted interactions.

11. A computer-implemented method comprising:

obtaining a plurality of images associated with an activity between a first person and a second person;

performing image analysis on the plurality of images to identify one or more body parts of the first person and one or more body parts of the second person;

using image analysis, tracking the one or more body parts of the first person and the one or more body parts of the second person over one or more images of the plurality of images;

based on the image analysis, detecting a physical contact between a given body part of the first person with a given body part of the second person in the one or more images of the plurality of images;

based on the image analysis, determining one or more positions of the given body part of the first person and the given body part of the second person at the physical contact in the one or more images of the plurality of images;

using the one or more positions, computing relative velocities of the given body part of the first person and the given body part of the second person at the physical contact;

based on the relative velocities and the one or more positions of the given body part of the first person and the given body part of the second person in the one or more images of the plurality of images, determining one or more characteristics of the physical contact; and

using a cognitive model, classifying the physical contact as a permitted interaction or an unpermitted interaction based at least on the one or more characteristics.

12. The method of claim 11 , wherein the plurality of images is obtained from a video.