DESCRIPTION PERSON TRACKING DEVICE, PERSON TRACKING METHOD, AND NON-TRANSITORY COMPUTER READABLE MEDIUM STORING PERSON TRACKING PROGRAM Technical Field [0001] The present invention relates to a person tracking device, a person tracking method, and a non-transitory computer readable medium storing a person tracking program and, particularly, to a person tracking device, a person tracking method, and a non-transitory computer readable medium storing a person tracking program that track a person using a video taken by a surveillance camera. Background Art [0002] Techniques to track a person using a video taken by a surveillance camera are disclosed in recent years. As one example of a person tracking method, Patent Literature 1 discloses a method of tracking a person based on a color feature of a person. [0003] Fig. 9 shows an exemplary embodiment of the person tracking system disclosed in Patent Literature 1. The person tracking system includes a person region extraction means 1, a voxel generation means 2, a person color feature extraction means 3, and a person tracking means 4. [0004] The person region extraction means 1 extracts a person region from a surveillance video and outputs a person region extraction result to the voxel generation means 2. The voxel generation means 2 generates voxel information from the person region extraction result output from the person region extraction means 1 and outputs the generated voxel information to the person color feature extraction means 3. The person color feature extracting means 3 extracts a person color feature from the voxel information output from the voxel generation means 2 and the surveillance video and outputs the extracted person color feature to the person tracking means 4. The person tracking means 4 tracks a person using the person color feature output from the person color feature extracting means 3 and outputs a person tracking result. [0005] The operation of the person tracking system shown in Fig. 9 is described in detail. [0006] The person region extraction means 1 extracts a person region from a surveillance video input from a camera using a background subtraction method. Then, the person region extraction means 1 outputs the extracted person region extraction result to the voxel generation means 2. [0007] The voxel generation means 2 generates voxels based on the input person region extraction result. The input person region extraction result is acquired by a plurality of cameras. The voxel generation means 2 projects the input person region extraction result onto the three-dimensional space using a volume intersection method and thereby generates voxels that represent the position of a person in the space. The voxel generation means 2 outputs the generated voxels to the person color feature extracting means 3. [0008] The person color feature extracting means 3 acquires the distribution of colors of a person from toe to tip in the vertical direction as a person color feature based on the generated voxels and the surveillance camera video. Specifically, the person color feature extracting means 3 calculates the average of colors for each height of the voxel, normalizes the result by height, and thereby calculates the person color feature. Although the color feature is basically determined by the color of clothes the person is wearing, the value obtained by calculating the average of colors in all directions at the same height is used. The person color feature extracting means 3 thereby achieves the extraction of the color feature that is robust against variation of the way the clothes look depending on the direction. [0009] The person tracking means 4 compares the obtained person color feature with a person color feature obtained in the past and thereby determines the similarity. The person tracking means 4 calculates the relationship between the voxels calculated in the past and the voxels calculated most recently in accordance with the determination result. Consequently, the person tracking means 4 calculates a person tracking result associating the past person extraction result and the current extraction result. Citation List Patent Literature [0010] PTL1: Japanese Unexamined Patent Application Publication No. 2005-250692 Summary of Invention Technical Problem [0011] In the person tracking system disclosed in Patent Literature 1, the tracking of a tracking target person is difficult when there is no distinctive feature in the clothes the tracking target person is wearing. The clothes are similar in general. When there are many persons who are wearing the clothes in the similar color to the clothes the tracking target person is wearing, the probability that the person tracking system confuses the tracking target person with another similar person increases, which makes accurate tracking difficult. Particularly, in the case of tracking a person using surveillance cameras with no overlap in their fields of view, once a tracking target person enters the blind spot of the camera and the tracking is discontinued temporarily, it is difficult for the person tracking system disclosed in Patent Literature 1 to correctly track the tracking target person even after the person comes back into the viewing range of the camera. [0012] The present invention has been accomplished to solve the above problems and an exemplary object of the present invention is thus to provide a person tracking device, a person tracking method, and a non-transitory computer readable medium storing a person tracking program that can achieve accurate tracking of a tracking target person even when the tracking target person has few distinctive features. Solution to Problem [0013] A person tracking device according to one aspect of the invention includes a person region information extraction means for detecting a person region where a person appearing in a video belongs, and generating person region information describing information of the person region; an accompanying person determination means for identifying at least one accompanying person accompanying a tracking target person among persons included in the person region information based on the person region information and information specifying a tracking target person, and generating accompanying person information describing the accompanying person; a distinctive person selection means for selecting a distinctive person having a salient feature using the person region information among the accompanying person specified by the accompanying person determination information, and generating distinctive person information describing the distinctive person; and a person tracking means for calculating a distinctive person tracking result being a tracking result for the distinctive person based on the person region information and the distinctive person information. [0014] A person tracking method according to one aspect of the invention includes detecting a person region where a person appearing in a video belongs, and generating information person region information describing information of the person region; identifying at least one accompanying person accompanying a tracking target person among persons included in the person region information based on the person region information and information specifying a tracking target person, and generating accompanying person information describing the accompanying person; selecting a distinctive person having a salient feature using the person region information among the accompanying person specified by the accompanying person determination information, and generating distinctive person information describing the distinctive person; and calculating a distinctive person tracking result being a tracking result for the distinctive person based on the person region information and the distinctive person information. [0015] A non-transitory computer readable medium storing a person tracking program according to one aspect of the invention is a non-transitory computer readable medium storing a program causing a computer to execute a process of tracking a person appearing in a video, the process including detecting a person region where the person appearing in the video belongs, and generating information person region information describing information of the person region; identifying at least one accompanying person accompanying a tracking target person among persons included in the person region information based on the person region information and information specifying a tracking target person, and generating accompanying person information describing the accompanying person; selecting a distinctive person having a salient feature using the person region information among the accompanying person specified by the accompanying person determination information, and generating distinctive person information describing the distinctive person; and calculating a distinctive person tracking result being a tracking result for the distinctive person based on the person region information and the distinctive person information Advantageous Effects of Invention [0016] According to the aspects of the invention, it is possible to provide a person tracking device, a person tracking method, and a non-transitory computer readable medium storing a person tracking program that can achieve accurate tracking of a tracking target person even when the tracking target person has few distinctive features. Brief Description of Drawings [0017] Fig. 1 is a block diagram showing a configuration of a person tracking device according to a first exemplary embodiment; Fig. 2 is a flowchart showing a flow of processing of an accompanying person determination unit 102 according to the first exemplary embodiment; Fig. 3 is a flowchart showing a flow of processing of the accompanying person determination unit 102 according to the first exemplary embodiment; Fig. 4 is a flowchart showing a flow of processing of a person tracking device 100 according to the first exemplary embodiment; Fig. 5 is a block diagram showing a configuration of a distinctive person selection unit 103 according to the first exemplary embodiment; Fig. 6 is a flowchart showing a flow of processing of a distinctive person determination unit 201 according to the first exemplary embodiment; Fig. 7 is a block diagram showing a configuration of a distinctive person selection unit 103 according to a second exemplary embodiment; Fig. 8 is a block diagram showing a configuration of a distinctive person selection unit 103 according to a third exemplary embodiment; and Fig. 9 is a block diagram showing a configuration of a person tracking system disclosed in Patent Literature 1. Description of Embodiments [0018]