Object recognition apparatus, classification tree learning apparatus and method thereof

PURPOSE: An object recognition apparatus, a classification tree leaning apparatus, and operation methods thereof are provided to recognize a plurality of components constituting an analysis target from one depth image of the analysis target without needing to use a plurality of depth images. CONSTITUTION: An input unit (1010) in which a depth image of an analysis target is inputted, and a processing unit (1020) recognizes a visible object part and a hidden object part of the analysis target from the depth image using a classification tree. A volume construction unit (1030) constructs a volume of the analysis target in one data space using the visible object part and the hidden object part. The processing unit extracts additional information of the analysis target based on the volume. The additional information includes a shape, a pose, a key joint, and a structure of the analysis target. [Reference numerals] (1010) Input unit; (1020) Processing unit; (1030) Volume construction unit; (1040) Size control unit