Systems and methods for 3D image distification
摘要
說明書
As described herein, the 2D image matrix 402, in some embodiments, may have a higher level of granularity with respect to the corresponding 2D-axis of raw 3D image 404. For example, a 2D point (not shown) on the 3D image 404 may exist within the rectangular space defined by, for example, points (X17, Y3), (X18, Y3), (X17, Y4) and (X18, Y4). Such a 2D point would have no direct mapping to the 2D image matrix 404. In such cases, when the 2D image matrix has fewer overall points than the raw 3D image, the 2D image matrix 404 is described as having a coarser granularity of 2D coordinates with respect to the available 2D coordinates of the 3D image. The courser granularity may occur because the image resolution (e.g., regarding the number of pixels) of the 3D image is higher than the number of 2D matrix points of the generated 2D image matrix 404. A coarser level of granularity for the 2D matrix 404 may be desirable in some embodiments, for example, in order to improve the performance of the computing device because fewer 2D coordinates of the 2D image matrix, compared to a greater number of such coordinates in the 3D image, could require less computing resources to process for certain applications, for example, the generation of a corresponding output feature vector, where the complexity of the corresponding output feature vector could depend on the level of granularity of the 2D image matrix. Thus, in some embodiments, coarser output feature vectors could likewise provide an improvement in further applications, such as when the output feature vectors are used to train or execute predictive models, as described herein.
