Sampling image records from a collection based on a change metric
View Patent ↗An image-record subject is identified for a plurality of image records in an image collection. Then, a sampling change-metric related to a changing characteristic of the image-record subject is identified, and sampling of at least a portion of the image collection occurs at least according to the sampling change-metric to obtain one or more image records for the subset. Information pertaining to results of the sampling step is stored in a computer-accessible memory system.
1. A method implemented at least in part by a computer system, the method for forming a subset of image records of a subject from at least a portion of an image collection based on growth of the subject, and the method comprising the steps of:
identifying an image-record subject for a plurality of image records in the image collection, wherein the subject is a person;
accessing a growth curve for a representative individual, wherein the growth curve represents age versus height;
determining sampling frequencies based at least upon the growth curve;
sampling at least a portion of the image collection using the sampling frequencies to obtain one or more image records for the subset of image records; and
storing information pertaining to results of the sampling step in a computer-accessible memory system.
2. The method of claim 1 , wherein the determining sampling frequencies step further includes using at least one different sampling criteria such as event classification.
3. The method of claim 2 , wherein the sampling frequencies for a first portion of the image collection corresponding to a high slope region of the growth curve are determined based upon the growth curve of a representative individual and the sampling frequencies for a second portion of the image collection corresponding to a low slope region of the growth curve are determined by a different sampling criteria such as event classification.
4. The method of claim 3 , wherein the sampling frequencies are determined as a function of the growth curve, event classification, and a sampling granularity factor.
5. The method of claim 2 , further comprising the step of clustering image records in at least a portion of the subset of image records based upon the sampling frequencies.
6. The method of claim 5 , wherein the clustering step uses a clustering granularity factor in addition to the sampling frequencies to cluster at least a portion of the subset of image records.
7. The method of claim 1 , further comprising the step of clustering image records in at least a portion of the subset of image records based upon the sampling frequencies.
8. The method of claim 1 , wherein the sampling frequencies are determined as a function of the growth curve and a sampling granularity factor.