Image driver that samples high-resolution image data
View Patent ↗Techniques are disclosed for generating a low-dimensional representation of an image. An image driver receives an image captured by a camera. The image includes features based on pixel values in the image, and each feature describes the image in one or more image regions. The image driver generates, for each of the plurality of features, a feature vector that includes values for that feature corresponding to at least one of the image regions. Each value indicates a degree that the feature is present in the image region. The image driver generates a sample vector from each of the feature vectors. The sample vector includes each of the values included in the generated feature vectors.
1. A computer-implemented method for generating a low-dimensional representation of an image, the method comprising:
receiving an image, wherein the image includes a plurality of features determined based on pixel values in the image and wherein each feature describes the image in one or more image regions;
generating, for each of the plurality of features, a feature vector that includes values for that feature corresponding to at least one of the image regions, wherein each value indicates a degree that the feature is present in the image region; and
generating a sample vector based on the generated feature vectors, wherein the sample vector includes each of the values included in the generated feature vectors, wherein the image is one in a sequence of images captured by at least one camera, and wherein the sample vector is evaluated relative to sample vectors corresponding to each image in the sequence.
2. The method of claim 1 , wherein the at least one camera is an infrared camera or a visible camera.
3. The method of claim 1 , further comprising:
sending the sample vector to a sample memory for retrieval by a machine learning engine that evaluates the sample vector.
4. The method of claim 1 , wherein the plurality of features includes at least one of a color feature, entropy feature, homogeneity feature, or gradient feature.
5. The method of claim 1 , further comprising, prior to receiving the image:
receiving a profile specifying a spatial granularity for at least one of the plurality of features;
adjusting an amount of regions to evaluate for the feature based on the specified spatial granularity.
6. A non-transitory computer-readable storage medium having instructions, which, when executed on at least one processor, cause the at least one processor to perform an operation for generating a low-dimensional representation of an image captured by a source device, the instructions comprising instructions to:
receive an image, wherein the image includes a plurality of features determined based on pixel values in the image and wherein each feature describes the image in one or more image regions,
generate, for each of the plurality of features, a feature vector that includes values for that feature corresponding to at least one of the image regions, wherein each value indicates a degree that the feature is present in the image region, and
generate a sample vector from the generated feature vectors, wherein the sample vector includes each of the values included in the generated feature vectors, wherein the image is one in a sequence of images captured by at least one source device, wherein the sample vector is evaluated relative to sample vectors corresponding to each image in the sequence.
7. The computer-readable storage medium of claim 6 , wherein the at least one source device includes an infrared camera or a visible camera.
8. The computer-readable storage medium of claim 6 , wherein the instructions further comprise instructions to:
send the sample vector to a sample memory for retrieval by a machine learning engine that evaluates the sample vector.
9. The computer-readable storage medium of claim 6 , wherein the plurality of features includes at least one of a color feature, entropy feature, homogeneity feature, or gradient feature.
10. The computer-readable storage medium of claim 6 , wherein the instructions further comprise instructions to, prior to receiving the image:
receive a profile specifying a spatial granularity for at least one of the plurality of features;
adjust an amount of regions to evaluate for the feature based on the specified spatial granularity.
11. A system, comprising:
a video input source configured to provide a sequence of video frames, each depicting a scene;
a central processing unit (CPU);
a graphics processing unit (GPU); and
a memory containing a program, which, when executed on the CPU and GPU, is configured to perform an operation for generating a low-dimensional representation of an image captured by at least one source device, the operation comprising:
receiving an image, wherein the image includes a plurality of features determined based on pixel values in the image and wherein each feature describes the image in one or more image regions,
generating, for each of the plurality of features, a feature vector that includes values for that feature corresponding to at least one of the image regions, wherein each value indicates a degree that the feature is present in the image region, and
generating a sample vector from each of the generated feature vectors, wherein the sample vector includes each of the values included in the generated feature vectors, wherein the image is one in a sequence of images captured by the at least one source device, wherein the sample vector is evaluated relative to sample vectors corresponding to each image in the sequence.
12. The system of claim 11 , wherein the at least one source device includes a camera.
13. The system of claim 12 , wherein the camera is an infrared camera or a visible camera.
14. The system of claim 11 , wherein the operation further comprises:
sending the sample vector to a sample memory for retrieval by a machine learning engine that evaluates the sample vector.
15. The system of claim 11 , wherein the operation further comprises, prior to receiving the image:
receiving a profile specifying a spatial granularity for at least one of the plurality of features;
adjusting an amount of regions to evaluate for the feature based on the specified spatial granularity.