Depth and motion estimations in machine learning environments

1 . At least one non-transitory computer-readable medium comprising instructions which, when executed, cause a computing device to perform operations comprising:

predicating values relating to a frame associated with a scene;

receiving the frame such that the values are associated with multiple loss functions corresponding to the frame, wherein the values are simultaneously predicted and comprise two or more of pixel depth, pixel velocity, pixel class and segmentation, or pixel optical flow; and

estimating one or more of depth or motion based on the predicted values.

2 . The non-transitory computer-readable medium of claim 1 , wherein the operations further comprise receiving and processing, by a neural network architecture, one or more of one or more previous frames for convolutional long short-term memory (LSTM) and odometry-based translation length, wherein the neural network architecture comprises a deep recurrent neural network architecture based on one or more deep recurrent neural networks or one or more convolutional LSTM layers.

3 . The non-transitory computer-readable medium of claim 1 , wherein the operations further comprise fusing together the multiple loss functions associated with the simultaneously predicted values in a cumulated cost function.

4 . The non-transitory computer-readable medium of claim 1 , wherein the operations further comprise estimating rotation matrixes and translation vectors for an object in the scene or one or more cameras capturing of the scene, wherein rotation matrixes and the translation vectors are fused together in a supervised form.

5 . The non-transitory computer-readable medium of claim 4 , wherein the operations further comprise constraining, based on odometry information, one or more of the rotation matrixes, translation vectors, and the simultaneously predicted values to estimate the depth of the scene and the motion of the one or more cameras, wherein the computing device comprises processing circuitry having one or more of graphics processing circuitry or application processing circuitry.

6 . A method comprising:

predicating, by processing circuitry of a computing device, values relating to a frame associated with a scene;

receiving the frame such that the values are associated with multiple loss functions corresponding to the frame, wherein the values are simultaneously predicted and comprise two or more of pixel depth, pixel velocity, pixel class and segmentation, or pixel optical flow; and

estimating one or more of depth or motion based on the predicted values.

7 . The method of claim 6 , further comprising receiving and processing, by a neural network architecture, one or more of one or more previous frames for convolutional long short-term memory (LSTM) and odometry-based translation length, wherein the neural network architecture comprises a deep recurrent neural network architecture based on one or more deep recurrent neural networks or one or more convolutional LSTM layers.

8 . The method of claim 6 , further comprising fusing together the multiple loss functions associated with the simultaneously predicted values in a cumulated cost function.

9 . The method of claim 6 , further comprising estimating rotation matrixes and translation vectors for an object in the scene or one or more cameras capturing of the scene, wherein rotation matrixes and the translation vectors are fused together in a supervised form.

10 . The method of claim 9 , further comprising constraining, based on odometry information, one or more of the rotation matrixes, translation vectors, and the simultaneously predicted values to estimate the depth of the scene and the motion of the one or more cameras, wherein the one or more cameras are coupled to the processing circuitry having one or more of graphics processing circuitry or application processing circuitry.

11 . An apparatus comprising:

processing circuitry to:

predict values relating to a frame associated with a scene; and

receive the frame such that the values are associated with multiple loss functions corresponding to the frame, wherein the values are simultaneously predicted and comprise two or more of pixel depth, pixel velocity, pixel class and segmentation, or pixel optical flow; and

estimate one or more of depth or motion based on the predicted values.

12 . The apparatus of claim 11 , wherein the processing circuitry is further to receive and process, via a neural network architecture, one or more of one or more previous frames for convolutional long short-term memory (LSTM) and odometry-based translation length, wherein the neural network architecture comprises a deep recurrent neural network architecture based on one or more deep recurrent neural networks or one or more convolutional LSTM layers.

13 . The apparatus of claim 12 , wherein the processing circuitry is further to estimate rotation matrixes and translation vectors for an object in the scene or one or more cameras capturing of the scene, wherein rotation matrixes and the translation vectors are fused together in a supervised form.

14 . The apparatus of claim 13 , wherein the processing circuitry is further to constrain, based on odometry information, one or more of the rotation matrixes, translation vectors, and the simultaneously predicted values to estimate the depth of the scene and the motion of the one or more cameras, wherein the one or more processors comprising one or more of a graphics processor or an application processor, wherein the one or more processors are co-located on a common semiconductor package.

15 . The apparatus of claim 11 , wherein the processing circuitry is further to fuse together the multiple loss functions associated with the simultaneously predicted values in a cumulated cost function.