Output File Format

XML Format

<?xml version="1.0" encoding="UTF-8"?>
<!--Root Element -->
<annotate>

    <!-- Set of Actions -->
    <actions>
        <action>A1</action>
         ...
    </actions>
    
    <!-- Video sequence -->
    <sequence frameCount="Fc" id="Cam" width="W" height="H" type="T">
        <!-- Frame -->
        <frame id="Fid" targetCount="N" filename="Fname">
            <target id="Tid" action="S">
                <location>x1,y1,x2,y2,...,xM,yM</location>
            </target>
			...
        </frame>
		...
    </sequence>
	...
				
    <!-- Matching of target in the video sequence -->
    <matching>
        <match fromSeq="mFS" fromID="mFID" toSeq="mTS" toID="mTOID">
				...
    </matching>
</annotate>

Definitions

Variable	Type	Definition
A1	string	Name of an action to be used in sequences
W	integer	Width of the frames in the sequence
H	integer	Height of the frames in the sequence
T	string	Type of sequence, e.g., "folder"
N	integer	Number of people detected in a particular frame
Fc	integer	Number of frames
targetCount	integer	Number of targets detected in the frame
Cam	integer	Camera index identifier, zero-based index
Fid	integer	Frame index identifier, zero-based index
S	string	Target state identifier from the action section
Fname	string	Filename of the frame, T must be "folder"
Tid	string	Target identifier
mFS	integer	Sequence identified number Cam value (from)
mTS	integer	Sequence identified number Cam value (to)
mFID	integer	Target id number (from)
mTID	integer	Target id number (to)
xi,yi	float	Coordinates (x,y) of the of the polygonal vertex number i

1. The list of actions describe the possible values of target's action. S = {A1, A2, A3, ... , AN}
2. State values are in camelBack format.
3. Target identifier is unique in all the sequence.
4. A match record can link two unique target's id from the same or different sequences.

Note about the states to be annotated (Example of Action states)

The states symbolized by variable S represents the following actions to be annotated by the user:

1. Line Up: A client (target) enters into this state the first time appears in the cashier's camera; it leaves this state when the same target moves to the cashier's area to place an order.
2. Order: A target enters into this state when it moves into the cashier's area to place an order; it leaves this state once it starts to pay.
3. Pay: A target changes enters this state when start paying for its order; it leaves this state when its starts moving away from the cashier's area.
4. Waiting: A target that moves away from the cashier area after paying enters into this state; it remains in this state despite in which camera appears until it starts picking up its order.
5. Pick Up: This state marks when the target picks it order from the picking area; it ends when the target moves away from the picking area with its order.
6. Leaving: This state marks the instant when the target moves away from the picking area with its order; it will remain in this state until it disappear from the view in all the cameras.

Status encoding

Status	Code
Line Up	lineUp
Order	order
Pay	pay
Waiting	wait
Pick Up	pickUp
Leave	leave

Example:

<annotate>
    <actions>
        <action>lineUp</action>
        <action>order</action>
        <action>pay</action>
        <action>pickUp</action>
        <action>leave</action>
    </actions>
    <sequence frameCount="3" id="0" width="640" height="360" type="folder">
        <frame id="0" targetCount="1" filename="00000001.jpg">
            <target id="0" action="null">
                <location>336,156,365,156,365,218,336,218</location>
            </target>
        </frame>
        <frame id="1" targetCount="1" filename="00000002.jpg">
            <target id="0" action="null">
                <location>336,156,365,156,365,218,336,218</location>
            </target>
        </frame>
        <frame id="2" targetCount="1" filename="00000003.jpg">
            <target id="0" action="null">
                <location>336,156,365,156,365,218,336,218</location>
            </target>
        </frame>
    </sequence>
    <sequence frameCount="3" id="1" width="1080" height="720" type="folder">
        <frame id="0" targetCount="1" filename="00000001.jpg">
            <target id="0" action="null">
                <location>336,156,365,156,365,218,336,218</location>
            </target>
        </frame>
        <frame id="1" targetCount="1" filename="00000002.jpg">
            <target id="0" action="null">
                <location>336,156,365,156,365,218,336,218</location>
            </target>
        </frame>
        <frame id="2" targetCount="1" filename="00000003.jpg">
            <target id="0" action="null">
                <location>336,156,365,156,365,218,336,218</location>
            </target>
        </frame>
    </sequence>
    <matching>
        <match fromSeq="0" fromID="0" toSeq="0" toID="2"/>
        <match fromSeq="1" fromID="3" toSeq="1" toID="1"/>
    </matching>
    <annotate>