Controlled-environment facility video communications monitoring system
Video communications of residents of a controlled-environment facility are monitored to detect instances of inappropriate behavior within the video, such as nudity. Residents may participate in live video visitation sessions with non-residents. Video recordings are generated for each video feed of the live video visitation. During the live video visitation, indications of inappropriate behavior displayed in the live video are detected. If indications of inappropriate behavior are detected in the live video, the recorded video is annotated to specify the locations of the detected indications in the recording. Upon completion of the video visitation session, post-processing of the recorded video detects additional indications of inappropriate behavior. The additional indications may be detected at proximal locations to the annotations and may be detected based on sampling frames of the recording to detect changes in the occurrence of pixels having a skin tone color.
1. A system for monitoring video communications provided to a resident of a controlled-environment facility, the system comprising:
a communication system configured to:
host a live video visitation between the resident and one or more non-residents, the video visitation comprising a plurality of live video feeds; and
a video processing system configured to:
sample a plurality of frames of a first video feed of the plurality of video feeds to detect indications of inappropriate behavior displayed in the first video feed;
identify a target region in one or more of the plurality of frames, the target region comprising a plurality of pixels having a color depth corresponding to a skin tone;
detect one or more sampled frames having an increased number of pixels with a color depth corresponding to a skin tone when compared to a prior sampled frame; and
trigger a predetermined action when indications of the inappropriate behavior are detected in the first video feed, the inappropriate behavior comprising at least partial nudity.
2. The system of claim 1 , wherein the predetermined action is one or more of alerting controlled-environment facility staff, terminating the first video feed, and marking the first video feed with an indication that inappropriate behavior is being detected.
3. The system of claim 1 , further comprising:
a video feed recording system configured to generate a plurality of video recordings, each video recording corresponding to a video feed of the plurality of live video feeds, and
wherein the video processing system is further configured to:
annotate a first video recording corresponding to the first video feed, wherein the annotations specify the locations of the detected indications of the inappropriate behavior in the first video recording.
4. The system of claim 1 , wherein the resident is associated with a privilege classification by the controlled-environment facility, and wherein the frames are sampled at a frequency determined based on the privilege classification of the resident.
5. The system of claim 4 , wherein a first sampling frequency is utilized when the resident is associated with a lower-risk privilege classification and wherein a second sampling frequency is utilized when the resident is associated with a higher-risk privilege classification, and wherein the second sampling frequency is higher than the first sampling frequency.
6. The system of claim 4 , wherein the sampling frequency is further determined based on a state of resource exhaustion by the video processing system.
7. The system of claim 1 , wherein the video processing system is further configured to modify the sampling rate based on the identification of the target region.
8. The system of claim 1 , wherein indications of inappropriate behavior are detected based upon the identification of predetermined forms or shapes in one or more frames of the first video feed.
9. A method for monitoring video communications of a resident of a controlled-environment facility, the method comprising:
hosting a live video visitation between the resident and one or more non-residents, the video visitation comprising a plurality of live video feeds, and wherein the resident is associated with a privilege classification by the controlled-environment facility;
generating a plurality of video recordings, each video recording corresponding to a live video feed of the plurality of live video feeds;
sampling a plurality of frames of a first recorded video feed to detect indications of inappropriate behavior displayed in the first video feed, the frames sampled at a frequency determined based on the privilege classification of the resident;
determining a color balance for a plurality of the frames sampled from the first recorded video feed;
determining a color balance baseline for the first recorded video feed based on the color balances determined for the plurality of the frames; and
annotating the first video recording, the annotations specifying the locations of the detected indications of the inappropriate behavior displayed in the first video recording.
10. The method of claim 9 , wherein a first sampling frequency is utilized when the resident is associated with a lower-risk privilege classification and wherein a second sampling frequency is utilized when the resident is associated with a higher-risk privilege classification, and wherein the second sampling frequency is higher than the first sampling frequency.
11. The method of claim 9 , wherein the sampling frequency is further determined based on a state of resource exhaustion by a video processing system.
12. The method of claim 9 , further comprising:
identifying a target region in one or more of the plurality of frames, wherein the target region comprises a plurality of pixels having a color depth corresponding to a skin tone.
13. The method of claim 12 , further comprising:
detecting one or more sampled frames having an increased number of pixels with a color depth corresponding to a skin tone when compared to a prior sampled frame.
14. The method of claim 12 , further comprising:
modifying the sampling rate based on the identification of the target region.
15. A method for monitoring video communications of a resident of a controlled-environment facility comprising:
hosting a live video visitation between the resident and one or more non-residents, the video visitation comprising a live video feed;
generating a video recording corresponding to the live video feed;
sampling a plurality of frames of a recorded video feed to detect indications of inappropriate behavior displayed in the video feed;
identifying a target region in one or more of the plurality of frames, the target region comprising a plurality of pixels having a color depth corresponding to a skin tone; and
annotating the video recording, the annotation specifying the locations of the detected inappropriate behavior displayed in the first video recording.
16. The method of claim 15 , wherein the resident is associated with a privilege classification by the controlled-environment facility, and wherein the frames are sampled at a frequency determined based on the privilege classification of the resident.
17. The method of claim 15 , wherein a first sampling frequency is utilized when the resident is associated with a lower-risk privilege classification and wherein a second sampling frequency is utilized when the resident is associated with a higher-risk privilege classification, and wherein the second sampling frequency is higher than the first sampling frequency.
18. The method of claim 15 , further comprising detecting one or more sampled frames having an increased number of pixels with a color depth corresponding to a skin tone when compared to a prior sampled frame.