WIPO logo
Mobile | Deutsch | English | Español | 日本語 | 한국어 | Português | Русский | 中文 | العربية |
PATENTSCOPE

Recherche dans les collections de brevets nationales et internationales
World Intellectual Property Organization
Recherche
 
Options de navigation
 
Traduction
 
Options
 
Quoi de neuf
 
Connexion
 
Aide
 
maximize
Traduction automatique
1. (WO2002045060) AMELIORATION DE LA RESOLUTION PAR FILTRAGE DE LA CLASSE VOISINE LA PLUS PROCHE
Note: Texte fondé sur des processus automatiques de reconnaissance optique de caractères. Seule la version PDF a une valeur juridique

CLAIMS
What is claimed is:
1. A method comprising:
enhancing image resolution by determining a nearest neighbor class for an input image vector from a plurality of spatial classes; and
applying a filter corresponding to the nearest neighbor class to the input image vector.

2. The method of claim 1 wherein determining the nearest neighbor class further comprises:
receiving the input image vector to be classified, and receiving a plurality of normalized mean class vectors, each class vector corresponding to a spatial classification;
normalizing the image vector;
determining weighted distances from the normalized image vector to each normalized mean class vector; and
determining which class vector is the nearest neighbor class to the input image vector based on the weighted distances.

3. The method of claim 2 wherein normalizing the input image vector further comprises:
determining a weighted mean for the input image vector.

4. The method of claim 3 wherein normalizing the input image vector further comprises:
determining a standard deviation for the input image vector.

5. An apparatus comprising: means for enhancing image resolution, including
means for determining a nearest neighbor class for an input image vector from a plurality of spatial classes; and
means for applying a filter corresponding to the nearest neighbor class to the input image vector.

6. The apparatus of claim 5 wherein said means for determining the nearest neighbor class further comprises:
means for receiving the input image vector to be classified, and for receiving a plurality of normalized mean class vectors, each class vector corresponding to a spatial classification;
means for normalizing the image vector;
means for determining weighted distances from the normalized image vector to each normalized mean class vector; and
means for determining which class vector is the nearest neighbor class to the input image vector based on the weighted distances.

7. The apparatus of claim 6 wherein said means for normalizing the input image vector further comprises:
means for determining a weighted mean for the input image vector.

8. . The apparatus of claim 7 wherein said means for normalizing the input image vector further comprises:
means for determining a standard deviation for the input image vector.

9. The apparatus of claim 9 further comprising:
means for adding the input image vector to the nearest neighbor class.

10. A computer readable medium having instructions which, when executed by a processing system, cause the system to:
enhance image resolution by determining a nearest neighbor class for an input image vector from a plurality of spatial classes; and
apply a filter corresponding to the nearest neighbor class to the input image vector.

11. _The medium of claim 10 wherein determining the nearest neighbor class further comprises:
receiving the input image vector to be classified, and receiving a plurality of normalized mean class vectors, each class vector corresponding to a spatial classification;
normalizing the image vector;
determining weighted distances from the normalized image vector to each normalized mean class vector;
determining which class vector is the nearest neighbor class to the input image vector based on the weighted distances.

12. The medium of claim 11 wherein normalizing the input image vector further comprises:
determining a weighted mean for the input image vector.

13. The medium of claim 12 wherein normalizing the input image vector further comprises:
determining a standard deviation for the input image vector.