Neural network-based image processing
A method and system for a deep learning-based approach to image processing to increase a level of optical zooming and increasing the resolution associated with a captured image. The system includes an image capture device to generate a display of a field of view (e.g., of a scene within a viewable range of a lens of the image capture device). An indication of a desired zoom level (e.g., 1.1× to 5×) is received, and, based on this selection, a portion of the field of view is cropped. In one embodiment, the cropped portion displayed by the image capture device for a user's inspection, prior to the capturing of a low resolution image. The low resolution image is provided to an artificial neural network trained to apply a resolution up-scaling model to transform the low resolution image to a high resolution image of the cropped portion.
1. A method comprising:
receiving, by an image capture device, an indication of a selected zoom level corresponding to a field of view of the image capture device;
cropping a first cropped portion of the field of view, the first cropped portion corresponding to the selected zoom level;
up-scaling the first cropped portion to generate an up-scaled version of the first cropped portion;
displaying, via an interface of the image capture device, the up-scaled version of the first cropped portion;
generating, in response to an image capture action, a first image corresponding to the first cropped portion, wherein the first image is generated at a first resolution, wherein the first image comprises a first image quality;
receiving, by an artificial neural network, the first image at the first resolution;
up-scaling, by the artificial neural network, the first resolution to a second resolution, wherein the second resolution is greater than the first resolution; and
generating, by the artificial neural network, a second image comprising the first cropped portion, wherein the second image is generated at the second resolution, wherein the second image comprises a second image quality, and wherein the second image quality is greater than the first image quality.
2. The method of claim 1 , further comprising:
applying, by the artificial neural network, one or more resolution up-scaling models to the first image.
3. The method of claim 1 , wherein the second resolution is greater than or equal to 300 dots per inch (dpi).
4. The method of claim 1 , wherein the image capture device is in a first preset shooting mode.
5. The method of claim 4 , wherein the artificial neural network applies a resolution up-scaling model corresponding to the first preset shooting mode.
6. A device comprising:
a memory to store instructions associated with an artificial neural network;
a sensor to capture a first image corresponding to a field of view;
a processing device operatively coupled to the memory and the sensor, the processing device to:
generate a view of a cropped portion of the field of view, the cropped portion corresponding to a selected zoom level;
display the view of the cropped portion at a first resolution;
generate, in response to an image capture action, the first image corresponding to the cropped portion at the first resolution; and
generate, by the artificial neural network, a second image of the cropped portion at a second resolution, wherein the second resolution is greater than the first resolution.
7. The device of claim 6 , the artificial neural network to apply a resolution up-scaling model to transform the first resolution to the second resolution.
8. The device of claim 6 , wherein the view of the cropped portion is displayed via a display of the device prior to execution of the image capture action.
9. The device of claim 6 , the processing device to receive an indication of the selected zoom level via an interface of the device.
10. The device of claim 6 , wherein the selected zoom level is a positive non-integer value.
11. The device of claim 6 , the processing device to:
operate in a feedback mode;
generate an image pair comprising the first image and the second image;
generate first data identifying at least one of a type of camera lens associated with the device or a shooting mode associated with the first image; and
provide the image pair and the first data to an operatively coupled service to cause an update to a resolution up-scaling model.
12. The device of claim 6 , wherein the first image is captured in a first preset shooting mode.
13. The device of claim 12 , wherein the artificial neural network applies a resolution up-scaling model corresponding to the first preset shooting mode.
14. The device of claim 6 , the processing device to up-scale the view of the cropped portion to the first resolution prior to display.
15. The device of claim 6 , wherein the artificial neural network is trained to comprise a plurality of resolution up-scaling models.
16. A non-transitory computer-readable storage device storing computer-executable instructions that, if executed by a processing device, cause the processing device to:
store data indicating a selected zoom level corresponding to a field of view of an image capture device;
cause a display of a cropped portion of the field of view, the cropped portion corresponding to the selected zoom level;
generate, in response to an image capture action, a first image corresponding to the field of view at a first image quality;
provide the data indicating the selected zoom level and the first image to an artificial neural network; and
generate, by the artificial neural network, a second image of the cropped portion corresponding to the selected zoom level at a second image quality, wherein the second image quality is greater than the first image quality.
17. The non-transitory computer-readable storage device of claim 16 , the artificial neural network to crop the first image of the field of view in accordance with the data indicating the selected zoom level.
18. The non-transitory computer-readable storage device of claim 16 , the artificial neural network to apply a model to the first image to generate the second image at the second image quality.
19. The non-transitory computer-readable storage device of claim 16 , the processing device to receive the data indicating the selected zoom level via an interface of the device.
20. The non-transitory computer-readable storage device of claim 16 , wherein the selected zoom level is a positive non-integer value.