Lossy compression for images and signals by identifying regions with low density of features
A device, system, and method perform lossy compression for images and signals by identifying regions with a low density of features. The method performed at a sensor communicatively connected to a receiver includes capturing sensor data. The method includes selecting a position in the sensor data. The method includes determining a local entropy of the position based on an entropy operation that indicates a probability distribution of a plurality of available values. When the local entropy is below a predetermined threshold, the method includes applying a first pre-processing operation to the position that averages features included in the position. The method includes transmitting the pre-processed sensor data corresponding to the position to the receiver.
1. A method, comprising:
at a sensor communicatively connected to a receiver:
capturing sensor data;
selecting a position in the sensor data;
determining a local entropy of the position based on an entropy operation that indicates a probability distribution of a plurality of available values;
when the local entropy is below a predetermined threshold, applying a first pre-processing operation to the position that averages features included in the position;
wherein the local entropy being below the predetermined threshold indicates a low density of the features in the position; and
transmitting the pre-processed sensor data corresponding to the position to the receiver.
2. The method of claim 1 , wherein, when the local entropy is at least the predetermined threshold, the method further comprises:
applying a second pre-processing operation to the position that retains the features included in the position.
3. The method of claim 2 , wherein the local entropy being at least the predetermined threshold indicates a high density of the features in the position.
4. The method of claim 1 , wherein the predetermined threshold is based on a maximum theoretical entropy of the available values.
5. The method of claim 1 , wherein the first pre-processing operation further denoises the sensor data corresponding to the position.
6. The method of claim 1 , wherein the local entropy is determined based on a presence function that satisfies a first criteria of being continuous relative to position and a second criteria of the presence of one of the features increasing a presence probability of another one of the features that is substantially similar.
7. The method of claim 6 , wherein the first criteria is satisfied by determining a frequency for each of the available values in a relative location in the position.
8. The method of claim 1 , wherein the local entropy is determined based on a Shannon Entropy analysis.
9. The method of claim 8 , wherein the local entropy is determined based on a product of an occurrence function each of the available values and a window function corresponding to the selected position.
10. The method of claim 9 , wherein the local entropy is determined based on a displacement of a distribution of the available values within the position.
11. The method of claim 1 , wherein the sensor data is image data.
12. A sensor, comprising:
a capturing component configured to capture sensor data;
a processor selecting a position in the sensor data, the processor determining a local entropy of the position based on an entropy operation that indicates a probability distribution of a plurality of available values, when the local entropy is below a predetermined threshold, the processor applying a first pre-processing operation to the position that averages features included in the position, wherein the local entropy being below the predetermined threshold indicates a low density of the features in the position; and
a transceiver configured to transmit the pre-processed sensor data corresponding to the position to a receiver that is communicatively connected to the sensor.
13. The sensor of claim 12 , wherein, when the local entropy is at least the predetermined threshold, the processor applies a second pre-processing operation to the position that retains the features included in the position.
14. The sensor of claim 13 , wherein the local entropy being at least the predetermined threshold indicates a high density of the features in the position.
15. The sensor of claim 12 , wherein the predetermined threshold is based on a maximum theoretical entropy of the available values.
16. The sensor of claim 12 , wherein the first pre-processing operation further denoises the sensor data corresponding to the position.
17. The sensor of claim 12 , wherein the sensor data is image data.
18. An imaging arrangement, comprising:
a receiver; and
a sensor capturing image data, the sensor selecting a position in the image data, the sensor determining a local entropy of the position based on an entropy operation that indicates a probability distribution of a plurality of available values, when the local entropy is below a predetermined threshold, the sensor applying a first pre-processing operation to the position that averages features included in the position, wherein the local entropy being below the predetermined threshold indicates a low density of the features in the position, when the local entropy is at least the predetermined threshold, the sensor applying a second pre-processing operation to the position that retains the features included in the position, the sensor transmitting the pre-processed sensor data corresponding to the position to the receiver,
wherein the receiver performs a compression operation on the pre-processed sensor data corresponding to the position.