Method and device for transforming 2D image into 3D
A method and device for transforming 2D images into 3D are disclosed. The disclosed device includes a dictionary storage unit configured to store a word-depth gradient dictionary; a color patch obtainer unit configured to obtain color patches from an input image; a matching word search unit configured to transform each of the color patches obtained by the color patch obtainer unit into a SIFT descriptor form and search for words closest to the SIFT descriptors of the obtained color patches from among the words of the word-depth gradient dictionary; a matching depth gradient obtainer unit configured to obtain depth gradient information of the words matching the obtained color patches from the word-depth gradient dictionary; and a depth map generation unit configured to compute a depth from the obtained matching depth gradient for each of the obtained color patches and generate a depth map.
1. A device for transforming a 2D image into 3D, the device comprising:
a processor; and
a memory storing one or more programs configured to be executed by the processor, the one or more programs comprising instructions for:
(a) storing a word-depth gradient dictionary, the word-depth dictionary having recorded therein words having a SIFT descriptor form and depth gradient information relating to each of the words;
(b) obtaining a plurality of color patches from an input image;
(c) transforming each of the plurality of color patches obtained by said step (b) into a SIFT descriptor form and searching for words closest to the SIFT descriptors of the obtained color patches from among the words of the word-depth gradient dictionary;
(d) obtaining depth gradient information of the words matching the obtained color patches from the word-depth gradient dictionary; and
(e) generating a depth map by computing a depth from the obtained matching depth gradient for each of the obtained color patches, wherein said step (d) comprises a step of detecting an edge area of the input image and said (d) obtains the color patches from the detected edge area.
2. The device for transforming a 2D image into 3D according to claim 1 , wherein the word-depth gradient dictionary is generated by way of a training process performed on a plurality of training images.
3. The device for transforming a 2D image into 3D according to claim 1 , wherein the words of the word-depth gradient dictionary are selected from resultant data resulting from transforming color patches obtained from training images into SIFT descriptors.
4. The device for transforming a 2D image into 3D according to claim 3 , wherein the words are selected from clusters after performing clustering on the resultant data resulting from transforming the color patches obtained from the training images into SIFT descriptors.
5. The device for transforming a 2D image into 3D according to claim 4 , wherein the words are SIFT descriptors positioned at centers of the clusters.
6. A method for transforming a 2D image into 3D, the method comprising:
(a) storing a word-depth gradient dictionary, the word-depth gradient dictionary having recorded therein words having a SIFT descriptor form and depth gradient information relating to each of the words;
(b) obtaining a plurality of color patches from an input image;
(c) transforming each of the plurality of color patches obtained by said step (b) into a SIFT descriptor form and searching for words closest to the SIFT descriptors of the obtained color patches from among the words of the word-depth gradient dictionary;
(d) obtaining depth gradient information of the words matching the obtained color patches from the word-depth gradient dictionary; and
(e) generating a depth map by computing a depth from the obtained matching depth gradient for each of the obtained color patches, wherein said step (d) comprises a step of detecting an edge area of the input image and said step (d) obtains the color patches from the detected edge area.
7. The method for transforming a 2D image into 3D according to claim 6 , wherein the word-depth gradient dictionary is generated by way of a training process performed on a plurality of training images.
8. The method for transforming a 2D image into 3D according to claim 6 , wherein the words are selected from clusters after performing clustering on the resultant data resulting from transforming the color patches obtained from the training images into SIFT descriptors.
9. The method for transforming a 2D image into 3D according to claim 8 , wherein the words are SIFT descriptors positioned at centers of the clusters.