Methods, systems, articles of manufacture and apparatus to reconstruct scenes using convolutional neural networks
Methods, apparatus, systems and articles of manufacture to reconstruct scenes using convolutional neural networks are disclosed. An example apparatus includes a sensor data acquirer to acquire ground truth data representing an environment, an environment detector to identify an environmental characteristic of the environment, a synthetic database builder to apply noise to the ground truth data to form a training set, a model builder to train a machine learning model using the training set and the ground truth data, and a model adjustor to modify the machine learning model to include residual OR-gate connections intermediate respective layers of the machine learning model. The synthetic database builder is further to store the machine learning model in association with the environmental characteristic of the environment.
1. An apparatus for generating a model for scene re-construction, the apparatus comprising:
interface circuitry;
machine-readable instructions; and
at least one processor circuit to be programmed by the machine-readable instructions to:
acquire ground truth data representing a first environment;
identify a first degree of optical reflectivity of the first environment;
apply noise to the ground truth data based on the first degree of optical reflectivity to form a training set;
train a machine learning model using the training set and the ground truth data;
modify the machine learning model to include residual OR-gate connections intermediate respective layers of the machine learning model;
store the machine learning model in association with the first degree of optical reflectivity of the first environment;
access sensor data;
identify a second degree of optical reflectivity of a second environment represented by the sensor data;
select the machine learning model based on the second degree of optical reflectivity matching the first degree of optical reflectivity; and
process the sensor data using the selected machine learning model to create a scene.
2. The apparatus of claim 1 , wherein one or more of the at least one processor circuit is to determine an error characteristic associated with the first degree of optical reflectivity, and apply the noise to the ground truth data based on the error characteristic.
3. The apparatus of claim 1 , wherein one or more of the at least one processor circuit is to split the ground truth data into a plurality of training sets and apply different amounts of noise to respective training sets, and train the machine learning model based on the plurality of training sets.
4. The apparatus of claim 3 , wherein one or more of the at least one processor circuit is to train the machine learning model using training sets with increasing amounts of noise.
5. The apparatus of claim 1 , wherein the machine learning model is implemented using a convolutional neural network (CNN).
6. The apparatus of claim 1 , wherein one or more of the at least one processor circuit is to identify the second degree of optical reflectivity based on metadata accessed in connection with the sensor data.
7. The apparatus of claim 1 , further including data storage to store the machine learning model.
8. At least one non-transitory computer readable medium comprising instructions to cause at least one processor circuit to at least:
acquire ground truth data representing a first environment;
identify a first degree of optical reflectivity of the first environment;
apply noise to the ground truth data based on the first degree of optical reflectivity to form a training set;
train a machine learning model using the training set and the ground truth data;
modify the machine learning model to include residual OR-gate connections intermediate respective layers of the machine learning model;
store the machine learning model in association with the first degree of optical reflectivity of the first environment;
access sensor data;
identify a second degree of optical reflectivity of a second environment represented by the sensor data;
select the machine learning model based on the second degree of optical reflectivity matching the first degree of optical reflectivity; and
process the sensor data using the selected machine learning model to create a scene.
9. The at least one non-transitory computer readable medium of claim 8 , wherein the instructions are to cause the at least one processor circuit to determine determining an error characteristic associated with the first degree of optical reflectivity, and wherein the noise applied to the ground truth data is based on the error characteristic.
10. The at least one non-transitory computer readable medium of claim 8 , wherein the instructions are to cause the at least one processor circuit to split the ground truth data into a plurality of training sets, apply different amounts of noise to respective ones of the plurality of the training sets, and train the machine learning model based on the plurality of training sets.
11. The at least one non-transitory computer readable medium of claim 10 , wherein the instructions are to cause the at least one processor circuit to train the machine learning model using training sets with increasing amounts of noise.
12. The at least one non-transitory computer readable medium of claim 8 , wherein the machine learning model is implemented using a convolutional neural network (CNN).
13. The at least one non-transitory computer readable medium of claim 8 , wherein the instructions are to cause the at least one processor circuit to identify the second degree of optical reflectivity based on metadata accessed in connection with the sensor data.
14. A method for generating models for scene re-construction, the method comprising:
acquiring ground truth data representing a first environment;
identifying a first amount of optical reflectivity of the first environment;
applying noise to the ground truth data based on the first amount of optical reflectivity to form a training set;
training a machine learning model using the training set and the ground truth data;
modifying the machine learning model to include residual OR-gate connections intermediate respective layers of the machine learning model;
storing the machine learning model in association with the first amount of optical reflectivity of the first environment;
accessing sensor data;
identifying a second amount of optical reflectivity of a second environment represented by the sensor data;
selecting the machine learning model based on the second amount of optical reflectivity matching the first amount of optical reflectivity; and
processing the sensor data using the selected machine learning model to create a scene.
15. The method of claim 14 , further including determining an error characteristic associated with the first amount of optical reflectivity, and wherein the noise applied to the ground truth data is based on the error characteristic.
16. The method of claim 14 , further including splitting the ground truth data into a plurality of training sets, wherein the applying of the noise to the training sets includes applying different amounts of noise to the respective training sets, and the training of the machine learning model is performed based on the plurality of training sets.
17. The method of claim 16 , wherein the training of the machine learning model is performed using training sets with increasing amounts of noise.
18. The method of claim 14 , wherein the machine learning model is implemented using a convolutional neural network (CNN).