3D point cloud enhancement with multiple measurements
Systems and methods are described for refining first point cloud data using at least second point cloud data and one or more sets of quantizer shifts. An example point cloud decoding method includes obtaining data representing at least a first point cloud and a second point cloud; obtaining information identifying at least a first set of quantizer shifts associated with the first point cloud; and obtaining refined point cloud data based on at least the first point cloud, the first set of quantizer shifts, and the second point cloud. The obtaining of the refined point cloud data may include performing a subtraction based on at least the first set of quantizer shifts. Corresponding encoding systems and methods are also described.
1. A point cloud decoding method comprising:
obtaining data representing at least a first point position in a first point cloud and a second point position in a second point cloud, the first point cloud representing a first observation of a scene and the second point cloud representing a second observation of the scene;
obtaining information identifying at least a first set of quantizer shifts associated with the first point cloud; and
obtaining refined point cloud data including a refined point position based on at least the first point position, the first set of quantizer shifts, and the second point position.
2. A point cloud decoder apparatus comprising a processor configured to perform at least:
obtaining data representing at least a first point position in a first point cloud and a second point position in a second point cloud, the first point cloud representing a first observation of a scene and the second point cloud representing a second observation of the scene;
obtaining information identifying at least a first set of quantizer shifts associated with the first point cloud; and
obtaining refined point cloud data including a refined point position based on at least the first point position, the first set of quantizer shifts, and the second point position.
3. The method of claim 1 , wherein obtaining the refined point cloud data comprises performing a subtraction based on at least the first set of quantizer shifts.
4. The method of claim 1 , further comprising:
obtaining information identifying a second set of quantizer shifts associated with the second point cloud, the second set of quantizer shifts being different from the first set of quantizer shifts;
wherein obtaining the refined point cloud data is further based on the second set of quantizer shifts.
5. The method of claim 1 wherein the first point cloud represents a left view of a scene and the second point cloud represents a right view of the scene.
6. The method of claim 1 , wherein the first point cloud and the second point cloud are frames associated with different times.
7. The method of claim 1 , wherein the first set of quantizer shifts comprises at least a first shift associated with the first point position in the first point cloud and a different second shift associated with a position of a different point in the first point cloud.
8. The method of claim 1 , wherein the data representing the first point cloud comprises a first parameter (y 1 ) representing the first point position having a first quantization range, the data representing the second point cloud comprises a second parameter (y 2 ) representing the second point position having a second quantization range, and wherein obtaining the refined point cloud data comprises obtaining a third parameter ({tilde over (x)}) representing the refined point position in the quantization range of both the first parameter (y 1 ) and the second parameter (y 2 ).
9. The method of claim 1 , wherein obtaining the refined point cloud data (x l ) based on the first point cloud (y l ) and the second point cloud (y r ) comprises selecting the refined first point cloud (x l ) to substantially maximize a product of factors comprising one or more of the following factors:
a conditional probability Pr(y l |x l ) of the first point cloud (y l ) given the refined first point cloud (x l ),
a conditional probability Pr(y r |g(x l )) of the second point cloud (y r ) given an estimate g(x l ) of a second refined point cloud (x r ), where the estimate g(x l ) is based on the first refined point cloud (x l ),
a prior probability Pr(x l ) of the first refined point cloud data (x l ), and
a prior probability Pr(g(x l )) of the estimate g(x l ) of the second refined point cloud (x r ).
10. A point cloud encoding method comprising:
obtaining data representing at least a first point position in a first point cloud and a second point position in a second point cloud, the first point cloud representing a first observation of a scene and the second point cloud representing a second observation of the scene;
quantizing the first point position in the first point cloud and the second point position in the second point cloud, wherein quantizing the first point position in the first point cloud data comprises adding at least a first set of quantizer shifts to the first point position such that quantization bins of the first point position and the second point position are not aligned; and
encoding in a bitstream the quantized first and second point clouds.
11. A point cloud encoding apparatus comprising a processor configured to perform at least:
obtaining data representing at least a first point position in a first point cloud and a second point position in a second point cloud, the first point cloud representing a first observation of a scene and the second point cloud representing a second observation of the scene;
quantizing the first point position in the first point cloud and the second point position in the second point cloud, wherein quantizing the first point position in the first point cloud data comprises adding at least a first set of quantizer shifts to the first point position such that quantization bins of the first point position and the second point position are not aligned; and
encoding in a bitstream the quantized first and second point clouds.
12. The method of claim 10 , further comprising encoding in the bitstream information indicating the first set of quantizer shifts.
13. The method of claim 10 , wherein quantizing the first point position and the second point position includes adding a second set of quantizer shifts to the second point cloud, further comprising encoding in the bitstream information indicating the second set of quantizer shifts.
14. The method of claim 10 , wherein the first point cloud represents a left view of a scene and the second point cloud represents a right view of the scene.
15. The decoder apparatus of claim 2 , wherein obtaining the refined point cloud data comprises performing a subtraction based on at least the first set of quantizer shifts.
16. The decoder apparatus of claim 2 , wherein the first point cloud represents a left view of a scene and the second point cloud represents a right view of the scene.
17. The decoder apparatus of claim 2 , wherein the first point cloud and the second point cloud are frames associated with different times.
18. The encoding method of claim 10 , wherein the first point cloud and the second point cloud are frames associated with different times.
19. The encoding apparatus of claim 11 , wherein the first point cloud represents a left view of a scene and the second point cloud represents a right view of the scene.
20. The encoding apparatus of claim 11 , wherein the first point cloud and the second point cloud are frames associated with different times.