Loading...
Please wait, while we are loading the content...
Similar Documents
Illumination Estimation Device, Illumination Estimation Method, and Illumination Estimation Program
| Content Provider | The Lens |
|---|---|
| Related Links | https://www.lens.org/images/patent/EP/3013045/B1/EP_3013045_B1.pdf |
| Language | English |
| Publisher Date | 2021-07-07 |
| Access Restriction | Open |
| Alternative Title | Beleuchtungsschätzungsvorrichtung, Beleuchtungsschätzungsverfahren Und Beleuchtungsschätzungsprogramm Dispositif D'estimation D'éclairement, Procédé D'estimation D'éclairement Et Programme D'estimation D'éclairement |
| Content Type | Text |
| Resource Type | Patent |
| Date Applied | 2014-05-22 |
| Agent | Vossius & Partner Patentanwälte Rechtsanwälte Mbb |
| Applicant | Nec Corp |
| Application No. | 14814071 |
| Claim | An illumination estimation device, comprising: a target object area extraction unit (102) which extracts an object area from a multispectral image, the object area being an area, including specular reflection, of an object; a distribution characteristics estimation unit (103) which estimates information representing a subspace that approximates a distribution of observation values in a spectral space, as distribution characteristics, the observation values being values to be recorded at respective pixels of the multispectral image, and the distribution characteristics being characteristics of a spreading manner of the observation values in the object area within the spectral space; a dimension reduction unit (104) which selects one or more distribution characteristics to be used from among the distribution characteristics estimated by the distribution characteristics estimation unit (103), the number of selected distribution characteristics being equal to a reduced number of dimensions, the reduced number of dimensions being determined depending on a light environment at a time of capturing the multispectral image; and an illumination estimation unit (105) which estimates spectral characteristics of illumination that illuminates the object based on the subspace information to be represented by the one or more distribution characteristics selected by the dimension reduction unit (104); and an illumination information storage unit (106) which stores information representing parameters for use in configuring an illumination model, wherein the illumination estimation unit (105) estimates the spectral characteristics of the illumination based on the subspace information and the illumination model configured based on the information stored in the illumination information storage unit (106), characterized in that the illumination estimation unit (105) selects an illumination condition which is most analogous to the subspace information represented by the one or more distribution characteristics selected by the dimension reduction unit, from the illumination model configured based on the information stored in the illumination storage unit (106), and estimates the spectral characteristics of the illumination as spectral characteristics of illumination specified by the illumination condition. The illumination estimation device according to claim 1, wherein the distribution characteristics estimation unit (103) estimates a principal component vector in a set of a predetermined number of the observation values, as the distribution characteristics. The illumination estimation device according to claim 1, wherein the distribution characteristics estimation unit (103) estimates, as the distribution characteristics, vectors constituting the subspace, the subspace being obtained by approximating the distribution of the observation values in the spectral space as the subspace including an origin, and a number of the vectors corresponds to the reduced number of dimensions. A computer-implemented illumination estimation method, comprising: extracting an object area from a multispectral image, the object area being an area, including specular reflection, of an object; estimating information representing a subspace that approximates a distribution of observation values in a spectral space, as distribution characteristics, the observation values being values to be recorded at respective pixels of the multispectral image, and the distribution characteristics being characteristics of a spreading manner of the observation values in the object area within the spectral space; selecting one or more distribution characteristics to be used from among the estimated distribution characteristics, the number of selected distribution characteristics being equal to a reduced number of dimensions, the reduced number of dimensions being determined in advance depending on a light environment at a time of capturing the multispectral image; and estimating spectral characteristics of illumination that illuminates the object based on the subspace information to be represented by the selected one or more distribution characteristics; and storing in advance information representing parameters for use in configuring an illumination model, and wherein the estimating estimates the spectral characteristics of the illumination based on the subspace information and the illumination model configured based on the stored information characterized in that the estimating the spectral characteristics selects an illumination condition which is most analogous to the subspace information represented by the one or more distribution characteristics selected, from the illumination model configured based on the information stored, and estimates the spectral characteristics of the illumination as spectral characteristics of illumination specified by the illumination condition. The illumination estimation method according to claim 4, wherein the estimating the information estimates a principal component vector in a set of a predetermined number of the observation values, as the distribution characteristics. The illumination estimation method according to claim 4, wherein the estimating the information estimates, as the distribution characteristics, vectors constituting the subspace, the subspace being obtained by approximating the distribution of the observation values in the spectral space as the subspace including an origin. An illumination estimation program, which, when executed by a computer, causes the computer to execute: a processing of extracting an object area from a multispectral image, the object area being an area, including specular reflection, of an object; a processing of estimating information representing a subspace that approximates a distribution of observation values in a spectral space, as distribution characteristics, the observation values being values to be recorded at respective pixels of the multispectral image, and the distribution characteristics being characteristics of a spreading manner of the observation values in the object area within the spectral space; a processing of selecting one or more distribution characteristics to be used from among the distribution characteristics estimated in the processing of estimating the information, the number of selected distribution characteristics being equal to a reduced number of dimensions, the reduced number of dimensions being determined depending on a light environment at a time of capturing the multispectral image; and a processing of estimating spectral characteristics of illumination that illuminates the object based on the subspace information to be represented by the one or more distribution characteristics selected in the processing of selecting; and the program causes the computer comprising a illumination information storage unit (106) which stores information representing parameters for use in configuring an illumination model to estimate the spectral characteristics of the illumination based on the subspace information and the illumination model configured based on the information stored in the illumination information storage unit (106) in the illumination estimation process, characterized in that the processing of estimating the spectral characteristics selects an illumination condition which is most analogous to the subspace information represented by the one or more distribution characteristics selected by processing of selecting, from the illumination model configured based on the information stored in the illumination storage unit (106), and estimates the spectral characteristics of the illumination as spectral characteristics of illumination specified by the illumination condition. |
| CPC Classification | Investigating Or Analysing Materials By Determining Their Chemical Or Physical Properties IMAGE DATA PROCESSING OR GENERATION; IN GENERAL PICTORIAL COMMUNICATION; e.g. TELEVISION |
| Extended Family | 063-464-482-484-133 188-287-265-212-746 083-035-420-477-253 021-978-419-359-991 030-156-188-443-910 132-208-709-943-751 132-411-870-558-351 005-476-281-640-843 |
| Patent ID | 3013045 |
| Inventor/Author | Kaneko Eiji |
| IPC | G06T7/90 G01N21/27 |
| Status | Active |
| Simple Family | 063-464-482-484-133 188-287-265-212-746 083-035-420-477-253 021-978-419-359-991 030-156-188-443-910 132-411-870-558-351 132-208-709-943-751 005-476-281-640-843 |
| CPC (with Group) | G01N21/27 G06T7/90 G06T2207/10024 H04N23/11 H04N23/56 |
| Issuing Authority | European Patent Office (EPO) |
| Kind | Patent/Patent 1st level of publication/Inventor's certificate |
