Abstract: Representations of an object in an image generated by an imaging apparatus can comprise two or more separate sub-objects, producing a compound object. Compound objects can negatively affect the quality of object visualization and threat identification performance. As provided herein, a compound object can be separated into sub-objects. Three-dimensional image data of a potential compound object is projected into a two-dimensional manifold projection, and segmentation is performed on the two-dimensional manifold projection of the compound object to identify sub-objects. Once sub-objects are identified, the two-dimensional, segmented manifold projection is projected into three-dimensional space. A three-dimensional segmentation may then be performed to identify additional sub-objects of the compound object that were not identified by the two-dimensional segmentation.

Abstract: Representations of an object in an image generated by an imaging apparatus can comprise two or more separate sub-objects, producing a compound object. Compound objects can negatively affect the quality of object visualization and threat identification performance. As provided herein, a compound object can be separated into sub-objects. Three-dimensional image data of a potential compound object is projected into a two-dimensional manifold projection, and segmentation is performed on the two-dimensional manifold projection of the compound object to identify sub-objects. Once sub-objects are identified, the two-dimensional, segmented manifold projection is projected into three-dimensional space. A three-dimensional segmentation may then be performed to identify additional sub-objects of the compound object that were not identified by the two-dimensional segmentation.

Abstract: Representations of an object (110) in an image generated by an imaging apparatus (100) can comprise two or more separate sub-objects, producing a compound object (500). Compound objects can negatively affect the quality of object visualization and threat identification performance. As provided herein, a compound object (500) can be separated into sub-objects. Three-dimensional image data of a potential compound object (500) is projected into a two-dimensional manifold projection (504), and segmentation is performed on the two-dimensional manifold projection of the compound object to identify sub-objects. Once sub-objects are identified, the two-dimensional, segmented manifold projection (900) is projected into three-dimensional space (1104). A three-dimensional segmentation may then be performed to identify additional sub-objects of the compound object that were not identified by the two-dimensional segmentation.

Abstract: Potential threat items may be concealed inside objects, such as portable electronic devices, that are subject to imaging for example, at a security checkpoint. Data from an imaged object can be compared to pre-determined object data to determine a class for the imaged object. Further, an object can be identified inside a container (e.g., a laptop inside luggage). One-dimensional Eigen projections can be used to partition the imaged object into partitions, and feature vectors from the partitions and the object image data can be used to generate layout feature vectors. One or more layout feature vectors can be compared to training data for threat versus non-threat-containing items from the imaged object's class to determine if the imaged object contains a potential threat item.

Abstract: Representations of an object (110) in an image generated by an imaging apparatus (100) can comprise two or more separate sub-objects, producing a compound object (500). Compound objects can negatively affect the quality of object visualization and threat identification performance. As provided herein, a compound object (500) can be separated into sub-objects. Three-dimensional image data of a potential compound object (500) is projected into a two-dimensional manifold projection (504), and segmentation is performed on the two-dimensional manifold projection of the compound object to identify sub-objects. Once sub-objects are identified, the two-dimensional, segmented manifold projection (900) is projected into three-dimensional space (1104). A three-dimensional segmentation may then be performed to identify additional sub-objects of the compound object that were not identified by the two-dimensional segmentation.

Abstract: Potential threat items may be concealed inside objects, such as portable electronic devices, that are subject to imaging for example, at a security checkpoint. Data from an imaged object can be compared to pre-determined object data to determine a class for the imaged object. Further, an object can be identified inside a container (e.g., a laptop inside luggage). One-dimensional Eigen projections can be used to partition the imaged object into partitions, and feature vectors from the partitions and the object image data can be used to generate layout feature vectors. One or more layout feature vectors can be compared to training data for threat versus non-threat-containing items from the imaged object's class to determine if the imaged object contains a potential threat item.