Multi-channel high-quality depth estimation system to provide augmented and virtual realty features

1. A multi-channel convolution Neural network (CNN) based depth estimation system for a monocular camera, wherein the depth estimation system comprising:

a monocular depth estimation module, wherein the monocular depth estimation module comprising:

a predictor unit for predicting depth in a single image based on pre-stored images and one or more parameters of the monocular camera; and

an edge alignment quality unit for removing edge discontinuities in the depth of the image by introducing a semantic head during a training to consider semantic objects to generate an aligned image;

a depth map refinement module, wherein the depth map refinement module comprising:

a panoptic segmentation unit for applying one or more semantic labels to one or more portions of the aligned image to generate a segmented image; and

a dictionary unit for applying a depth pattern to each of the one or more portions of the segmented image based on the one or more semantic labels to generate a processed image;

a depth layout module, wherein the depth layout module facilitates the depth map refinement module by providing a depth pattern to one or more unrecognized semantic labels in the processed image to form a labelled image;

a depth inpainting module, wherein the depth inpainting module imprints one or more occluded regions in the labelled image to generate an inpainted image; and

an output module, wherein the output module adds a plurality of augmented reality and virtual reality features to the inpainted image to produce a three-dimensional image.

2. The depth estimation system in accordance with claim 1 , wherein the one or more parameters include focal length, principal centers within a prediction pipeline.

3. The depth estimation system in accordance with claim 2 , wherein the pre-stored images and the one or more parameters are imported in the prediction pipeline.

4. The depth estimation system in accordance with claim 3 , wherein the pre-stored images and the one or more parameters constitute a training data.

5. The depth estimation system in accordance with claim 1 , wherein the edge alignment quality unit introduces the semantic head for removing the plurality of edge discontinuities to create the aligned image.

6. The depth estimation system in accordance with claim 1 , wherein the panoptic segmentation unit generates the semantic labels for one or more portions with a depth pattern model.

7. The depth estimation system in accordance with claim 1 , wherein a depth map includes one or more vanishing points to indicate depth variation patterns.

8. The depth estimation system in accordance with claim 1 , wherein the one or more unrecognized labels are kept in an ordinal order.

9. The depth estimation system in accordance with claim 1 , wherein the depth inpainting module includes a semantic edge detector.

10. The depth estimation system in accordance with claim 9 , wherein the semantic edge detector detects a plurality of semantic edges.

11. The depth estimation system in accordance with claim 9 , wherein the semantic edge detector allows an increased inpainting speed.

12. The depth estimation system in accordance with claim 9 , wherein the semantic edge detector includes a fusion mechanism to detect the semantic edges based on occlusions.

13. The depth estimation system in accordance with claim 1 , wherein the depth map refinement module performs a local smoothness operation.

14. The depth estimation system in accordance with claim 1 , wherein the depth layout module performs as a global smoothness operation.

15. A method of depth estimation for a monocular camera based on multi-channel convolution Neural network (CNN), wherein the method comprising:

receiving a single image and improving a quality of depth within the image;

predicting a depth of the image on the basis of pre-stored images and one or more parameters;

removing a plurality of edge discontinuities in the depth by introducing a semantic head during a training to consider one or more semantic objects to create an aligned image;

performing refinement of the aligned image by obtaining one or more semantic labels for one or more portions of the aligned image to form a segmented image;

applying a depth pattern to each entity of the one or more portions of the aligned image based on the one or more semantic labels to generate a processed image;

providing one or more depth patterns to one or more unrecognized semantic labels in the processed image to form a labelled image;

inpainting one or more occluded regions in the labelled image creating an inpainted image;

reconstructing a three dimensional scene on the inpainted image; and

adding a plurality of augmented reality and virtual reality features to the inpainted image to generate a three dimensional image.

16. A depth estimation system for monocular camera, wherein the system comprising:

a monocular depth estimation module, wherein the monocular depth estimation unit comprising:

a predictor unit for predicting a depth of a single image on the basis of pre-stored images and one or more internal parameters, wherein the pre-stored images and the one or more internal parameters are imported into a prediction pipeline; and

an edge alignment quality unit for removing edge discontinuities in the depth by introducing a semantic head during a training to consider semantic objects to generate an aligned image;

a depth map refinement module, wherein the depth map refinement unit comprising:

a panoptic segmentation unit for obtaining one or more semantic labels for one or more portions of the aligned image to form a segmented image; and

a dictionary unit for applying a depth pattern to each portion of the one or more portions in the segmented image based on the one or more semantic labels to generate a processed image;

a depth layout module, wherein the depth layout module includes one or more vanishing points to indicate depth variation patterns to facilitate depth map refinement module for providing depth patterns to one or more unrecognized semantic labels in the processed image, further wherein one or more unrecognized labels are kept in a correct ordinal order to form a labelled image;

a depth inpainting module, wherein the depth inpainting module comprises a semantic edge detector for inpainting one or more occluded regions in the labelled image creating an inpainted image and removing one or more spurious depth edges to increase an inpainting speed; and

an output module, wherein the output module produces a three dimensional scene reconstruction of the inpainted image and adding a plurality of augmented reality and virtual reality features to the inpainted image to generate a three dimensional image.