| 摘要 |
The present invention relates to a system for generating a montage using facial features and a method for the same. The system comprises: a face type selection unit (100) which displays face sample images which are pre-stored in a montage DB (10) and which each have property values, and enables a witness to select face sample images similar to a suspect among the displayed face sample images; a face type synthesis unit (200) which synthesizes the face sample images selected by the face type selection unit (100) at the same ratio or a ratio inputted by the operation of the witness to create a first montage (F1); a face type adjustment unit (300) which indexes, from the montage DB (10), face sample images having property values similar to the property values of the first montage (F1) created by the face type synthesis unit (200), and synthesizes the indexed face sample images with the first montage (F1) to create a plurality of peripheral face images having property values similar to the property values of the first montage (F1); and a face type selection unit (400) which synthesizes a plurality of peripheral face images selected by the operation of the witness, among the peripheral face images created by the face type adjustment unit (300), at the same ratio or a ratio inputted by the operation of the witness to create a second montage (F2). According to the present invention as described above, the system proposes candidate groups relative to the face of a suspect using facial features according to hierarchical property statistics and enables a witness to select the candidate groups to create montages, thereby effectively generating montages even when the statement of the witness is not sufficient or the witness cannot express his memory easily in words. |
