Augmenting reliable training data with CycleGAN for hand pose estimation
A method and apparatus for generating augmented training data for hand pose estimation include receiving source data that is associated with a first lighting condition. Target data that is associated with a second lighting condition is received. A lighting condition translation between the first lighting condition and the second lighting condition is determined. Lighting translated data is generated based on the lighting condition translation and the source data. Augmented training data for hand pose estimation is generated based on the target data and the lighting translated data.
1. A method for generating augmented training data for hand pose estimation, comprising:
receiving, by a device, synthetic hand pose image that includes a first lighting condition;
receiving, by the device, a background image that includes a second lighting condition;
determining, by the device and by inputting only the synthetic hand pose image and the background image into a CycleGAN, a lighting condition translation that maps the first lighting condition and the second lighting condition;
generating, by the device, a lighting translated synthetic hand pose image using the synthetic hand pose image based on the lighting condition translation between the first lighting condition and the second lighting condition; and
generating, by the device, the augmented training data including the lighting translated synthetic hand pose image that is superimposed on the background image for hand pose estimation based on the background image and the synthetic hand pose image.
2. The method of claim 1 , further comprising:
training a hand pose estimation model using the augmented training data.
3. The method of claim 1 , further comprising:
generating, using a three dimensional model simulator, the synthetic hand pose image that represents a hand pose.
4. The method of claim 1 , wherein the background image is a real-world image.
5. A device, comprising:
at least one memory configured to store program code;
at least one processor configured to read the program code and operate as instructed by the program code, the program code including:
receiving code that is configured to cause the at least one processor to:
receive a synthetic hand pose image that includes a first lighting condition; and
receive a background image that includes a second lighting condition;
determining code that is configured to cause the at least one processor to determine, by inputting only the synthetic hand pose image and the background image into a CycleGAN, a lighting condition translation that maps the first lighting condition and the second lighting condition; and
generating code that is configured to cause the at least one processor to:
generate a lighting translated synthetic hand pose image using the synthetic hand pose image based on the lighting condition translation between the first lighting condition and the second lighting condition; and
generate augmented training data including the lighting translated synthetic hand pose image that is superimposed on the background image for hand pose estimation based on the background image and the lighting translated synthetic hand pose image.
6. The device of claim 5 , further comprising:
training code that is configured to cause the at least one processor to train a hand pose estimation model using the augmented training data.
7. The device of claim 5 , wherein the generating code is further configured to cause the at least one processor to generate, using a three dimensional model simulator, the synthetic hand pose image that represents a hand pose.
8. The device of claim 5 , wherein the background image is a real-world image.
9. A non-transitory computer-readable medium storing instructions, the instructions comprising: one or more instructions that, when executed by one or more processors of a device, cause the one or more processors to:
receive a synthetic hand pose image that includes a first lighting condition;
receive a background image that includes a second lighting condition;
determine, by inputting only the synthetic hand pose image and the background image into a CycleGAN, a lighting condition translation that maps the first lighting condition and the second lighting condition;
generate a lighting translated synthetic hand pose image using the synthetic hand pose image based on the lighting condition translation between the first lighting condition and the second lighting condition; and
generate augmented training data including the lighting translated synthetic hand pose image that is superimposed on the background image for hand pose estimation based on the background image and the lighting translated synthetic hand pose image.
10. The non-transitory computer-readable medium of claim 9 , wherein the one or more instructions cause the one or more processors to:
train a hand pose estimation model using the augmented training data.
11. The non-transitory computer-readable medium of claim 9 , wherein the background image is a real-world image.