Systems and methods for detecting egress at an entrance of a retail facility

Abstract

In some embodiments, apparatuses and methods are provided herein useful to detecting egress. In some embodiments, there is provided a system for detecting egress at an entrance including a video camera; a computer; a network; and a control circuit configured to receive live video footage; detect a human and estimate a location of the human; track a location and movement of a detected human; determine that the detected human has moved from a first region to a second region and to a third region; and transmit an alert message that indicates that the human has exited the retail facility through the entrance area. In some embodiments, the systems and methods may be configured to comply with privacy requirements which may vary between jurisdictions. For example, before any recording, collection, capturing or processing of user images (e.g., video image, video footage, etc.), a “consent to capture” process may be implemented.

What is claimed is: 
     
         1 . A system for use in detecting egress at an entrance of a retail facility, the system comprising:
 a video camera mounted and arranged to capture video footage of an entrance area of a retail facility in real time, wherein the entrance area is not intended to be an exit, wherein the video camera comprises a low resolution camera capturing images at less than 800 pixels by 600 pixels;   a computer at the retail facility and coupled to the video camera and configured to receive the video footage from the video camera;   a network coupled to the computer; and   a control circuit coupled to the network, wherein the control circuit is at a central location remote from the retail facility and is configured to receive the live video footage from the computer via the network, wherein the control circuit is configured to:
 detect, for each of a plurality of frames of the video footage and using a human detection module comprising a neural network model, a human and estimate a location of the human within at least one of a first region, a second region and a third region of the entrance area, wherein the first region corresponds to a region inside of a doorway of the entrance area, wherein the second region corresponds to a region proximate the doorway, and wherein the third region corresponds to a region outside of the doorway and further out the doorway than the second region, and wherein the neural network model is trained using stored footage from one of the video camera and a similar video camera having similar resolution and point of view; 
 track, over the plurality of frames of the video footage and using a human tracking module according to a set of rules, a location of a detected human, and movement of the detected human relative to the first region, the second region and the third region; 
 determine that the detected human has moved from the first region to the second region and to the third region; and 
 transmit an alert message that indicates that the human has exited the retail facility through the entrance area, wherein the alert message further comprises a camera identifier, a store identifier and an image capture of the detected human. 
   
     
     
         2 . The system of  claim 1 , wherein the at least one of the first region, the second region and the third region of the entrance area are formed from at least two preset horizontal thresholds applied across each of the plurality of frames, wherein the set of rules comprises the at least two preset horizontal thresholds determined based on the video camera and the entrance area. 
     
     
         3 . The system of  claim 1 , wherein the entrance area comprises an unguarded entrance. 
     
     
         4 . The system of  claim 1 , further comprising a tracking database coupled to the control circuit and configured to store a plurality of identifiers each associated with one of detected human and tracked human on each current frame that is currently being processed by the control circuit for each of a plurality of retail facilities. 
     
     
         5 . The system of  claim 1 , wherein the control circuit is further configured to:
 store the alert message over a period of time;   determine a total count of alert messages over the period of time;   determine whether the total count has reached a threshold value; and   in response to the total count reaching the threshold value, provide a notification message to an electronic device indicating an assignment of an associate at the entrance area.   
     
     
         6 . The system of  claim 1 , wherein the control circuit is further configured to:
 store occurrences of the alert message over a period of time;   determine occurrences of shrinkage at the retail facility over the period of time; and   correlate the occurrences of the alert message with the occurrences of shrinkage over the period of time.   
     
     
         7 . The system of  claim 1 , further comprising an electronic device configured to receive the alert message, wherein the electronic device is worn by an associate at the retail facility. 
     
     
         8 . The system of  claim 1 , wherein the alert message further comprises a timestamp corresponding to a time the human exited the retail facility through the entrance area. 
     
     
         9 . A method for use in counting humans at an area of a retail facility over a period of time, the method comprising:
 capturing, by a video camera mounted and arranged to capture, video footage of an entrance area of a retail facility in real time, wherein the video camera comprises a low resolution camera capturing images at less than 800 pixels by 600 pixels;   receiving, by a computer at the retail facility and coupled to the video camera via a network, video footage from the video camera;   receiving, by a control circuit coupled to the network and at a central location remote from the retail facility, the live video footage from the computer;   detecting a human and estimating, by the control circuit for each of a plurality of frames of the video footage and using a human detection module comprising a neural network model, a location of the human within at least one of a first region, a second region and a third region of the entrance area, wherein the first region corresponds to a region inside of a doorway of the entrance area, wherein the second region corresponds to a region proximate the doorway, and wherein the third region corresponds to a region outside of the doorway and further out the doorway than the second region, and wherein the neural network model is trained using stored footage from one of the video camera and a similar video camera having similar resolution and point of view;   tracking, by the control circuit, over the plurality of frames of the video footage and using a human tracking module according to a set of rules, a location of a detected human, and movement of the detected human relative to the first region, the second region and the third region;   determining, by the control circuit, that the detected human has moved from the first region to the second region and to the third region; and   transmitting, by the control circuit, an alert message that indicates that the human has exited the retail facility through the entrance area, wherein the alert message further comprises a camera identifier, a store identifier and an image capture of the detected human.   
     
     
         10 . The method of  claim 9 , wherein the at least one of the first region, the second region and the third region of the entrance area are formed from at least two preset horizontal thresholds applied across each of the plurality of frames, wherein the set of rules comprises the at least two preset horizontal thresholds determined based on the video camera and the entrance area. 
     
     
         11 . The method of  claim 9 , wherein the entrance area comprises an unguarded entrance. 
     
     
         12 . The method of  claim 9 , further comprising storing, by a tracking database coupled to the control circuit, a plurality of identifiers each associated with one of detected human and tracked human on each current frame that is currently being processed by the control circuit for each of a plurality of retail facility. 
     
     
         13 . The method of  claim 9 , further comprising:
 storing, by the control circuit, the alert message over a period of time;   determining, by the control circuit, a total count of alert messages over the period of time;   determining, by the control circuit, whether the total count has reached a threshold value; and   in response to the total count reaching the threshold value, providing, by the control circuit, a notification message to an electronic device indicating an assignment of an associate at the entrance area.   
     
     
         14 . The method of  claim 9 , further comprising:
 storing, by the control circuit, occurrences of the alert message over a period of time;   determining, by the control circuit, occurrences of shrinkage at the retail facility over the period of time; and   correlating, by the control circuit, the occurrences of the alert message with the occurrences of shrinkage over the period of time.   
     
     
         15 . The method of  claim 9 , further comprising receiving, by an electronic device coupled to the network, the alert message, wherein the electronic device is worn by an associate at the retail facility. 
     
     
         16 . The method of  claim 9 , wherein the alert message further comprises a timestamp corresponding to a time the human exited the retail facility through the entrance area. 
     
     
         17 . A system for use in counting humans at an area of a retail facility over a period of time, the system comprising:
 a video camera mounted and arranged to capture video footage of an area of a retail facility in real time, wherein the video camera comprises a low resolution camera capturing images at less than 800 pixels by 600 pixels;   a computer at the retail facility and coupled to the video camera and configured to receive the video footage from the video camera;   a network coupled to the computer; and   a control circuit coupled to the network, wherein the control circuit is at a central location remote from the retail facility and is configured to receive the live video footage from the computer via the network, wherein the control circuit is configured to:
 detect, for each of a plurality of frames of the video footage and using a human detection module comprising a neural network model, humans and estimate locations of the detected humans within the area, wherein the neural network model is trained using a database of stored footage from one of the video camera and a similar video camera having similar resolution and point of view; 
 assign an identifier to each detected human; 
 define a count as a total number of identified and the detected humans; 
 track, over the plurality of frames of the video footage spanning the period of time and using a human tracking module according to a set of rules, the location of each identified and detected human and movement of each identified and detected human within the area, wherein the control circuit is configured to track by:
 comparing detected humans in each frame to identified and detected humans from a previous frame;
 for detected humans having a similarity score corresponding to a previously identified and detected human, not incrementing the count; 
 for detected humans having a similarity score not corresponding to a previously identified and detected human, assigning a new identifier and incrementing the count; and 
 for previously identified and detected humans that do not have a similarity score corresponding to any of the detected humans of a current frame, decrement the count; and 
 
 
 transmit, at a conclusion of the plurality of frames, an alert message that indicates a value of the count, wherein the alert message further comprises a store identifier, a camera identifier, and an area identifier. 
   
     
     
         18 . A system for use in determining a count of people at a retail facility, the system comprising:
 a first video camera mounted and arranged to capture a first video footage of an entrance area of a retail facility in real time, wherein the entrance area is not intended to be an exit, wherein the first video camera comprises a first low resolution camera capturing images at less than 800 pixels by 600 pixels;   a second video camera mounted and arranged to capture a second video footage of an exit area of the retail facility in real time, wherein the exit area is not intended to be an entrance, wherein the second video camera comprises a second low resolution camera capturing images at less than 800 pixels by 600 pixels;   a computer at the retail facility and coupled to the first and second video cameras and configured to receive the first and second video footages from the first and second video cameras;   a network coupled to the computer; and   a control circuit coupled to the network, wherein the control circuit is at a central location remote from the retail facility and is configured to receive the live video footage from the computer via the network, wherein the control circuit is configured to:
 detect, for each of a plurality of frames of the first and second video footages and using a human detection module comprising a neural network model, one or more humans and estimate locations of the one or more humans within at least one of a first region, a second region and a third region of the entrance area and the exit area, wherein the first region corresponds to a region inside of a corresponding doorway of the entrance area and the exit area, wherein the second region corresponds to a region proximate the corresponding doorway, and wherein the third region corresponds to a region outside of the corresponding doorway and further out the corresponding doorway than the second region, and wherein the neural network model is trained using stored footage from one of the first video camera, the second video camera, and a similar video camera having similar resolution and point of view; 
 track, over the plurality of frames of the first and second video footages and using a human tracking module according to a set of rules, a location of a detected human, and movement of the detected human relative to the first region, the second region and the third region; 
 determine that the detected human has moved from the first region to the second region and to the third region and, in response, increment an outbound value; 
 determine that the detected human has moved from the third region to the second region and to the first region and, in response, increment an inbound value; 
 subtract the outbound value from the inbound value to determine a count of people inside the retail facility; 
 store the count over a period of time to a database coupled to the control circuit; and 
 cause an electronic device to display the count.