http://rdf.ncbi.nlm.nih.gov/pubchem/patent/CN-107301644-B
Outgoing Links
| Predicate | Object |
|---|---|
| classificationCPCAdditional | http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T2207-10024 |
| classificationCPCInventive | http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T7-136 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06F18-2321 http://rdf.ncbi.nlm.nih.gov/pubchem/patentcpc/G06T7-11 |
| classificationIPCInventive | http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06T7-136 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06T7-11 http://rdf.ncbi.nlm.nih.gov/pubchem/patentipc/G06K9-62 |
| filingDate | 2017-06-09-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
| grantDate | 2019-10-08-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
| publicationDate | 2019-10-08-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
| publicationNumber | CN-107301644-B |
| titleOfInvention | Unsupervised Segmentation of Natural Images Based on Mean Shift and Fuzzy Clustering |
| abstract | The invention is an unsupervised segmentation method of natural images based on mean shift and fuzzy clustering, which mainly solves the problem of low accuracy of unsupervised segmentation of massive natural images in the prior art. The scheme is: 1) input an image and smooth it; 2) uniformly initialize 64 iteration initial points in the normalized RGB color space of the smoothed image pixels; 3) perform iterative search on the initial points to obtain 64 converged 4) Delete the convergence points whose number of pixels in the high-dimensional sphere centered on the convergence point is less than the deletion threshold; 5) Merge the convergence points whose Euclidean distance is less than the merger threshold, determine the density peak and the number of density peaks, and calculate the membership degree and pixel 6) Defuzzify the smooth membership of the pixel, add a class label to each pixel, and output the segmented image. The invention does not need to set control parameters, can automatically determine the number of image segmentation categories, and can be used for unsupervised segmentation of massive natural images. |
| priorityDate | 2017-06-09-04:00^^<http://www.w3.org/2001/XMLSchema#date> |
| type | http://data.epo.org/linked-data/def/patent/Publication |
Incoming Links
| Predicate | Subject |
|---|---|
| isDiscussedBy | http://rdf.ncbi.nlm.nih.gov/pubchem/substance/SID425901710 http://rdf.ncbi.nlm.nih.gov/pubchem/compound/CID712 |
Total number of triples: 17.