Relevant motion detection in video
Methods, systems, and/or apparatuses are described for detecting relevant motion of objects of interest (e.g., persons and vehicles) in surveillance videos. As described herein input data based on a plurality of captured images and/or video is received. The input data may then be pre-processed and used as an input into a convolution network that may, in some instances, have elements that perform both spatial-wise max pooling and temporal-wise max pooling. Based on The convolution network may be used to generate a plurality of prediction results of relevant motion of the objects of interest.
1. A method comprising:
receiving, by a computing device, a sequence of images;
determining image features of the sequence of images based on:
applying a spatial filter that filters spatial data of the sequence of images; and
after applying the spatial filter, applying a sequential filter that filters sequential data of the sequence of images;
determining, based on the image features, that the sequence of images indicates motion of one or more objects; and
outputting, based on a query associated with the one or more objects, one or more of the images.
2. The method of claim 1 , wherein the spatial filter reduces the spatial data and the sequential filter reduces the sequential data.
3. The method of claim 1 , wherein the spatial filter reduces the spatial data based on at least one of:
reducing a quantity of data values, in each of a plurality of data groups associated with pixel dimensions of the images, based on data values of the data group; or
determining a data value, for each of a plurality of data groups associated with pixel dimensions of the images, based on a maximum of data values of the data group.
4. The method of claim 1 , wherein the sequential filter reduces the sequential data based on at least one of:
reducing a quantity of data values, in each of a plurality of data groups associated with the sequence of images, based on data values of the data group; or
determining a data value, for each of a plurality of data groups associated with the sequence of images, based on a maximum of data values of the data group.
5. The method of claim 1 , further comprising one or more of:
determining, before the determining the image features and by removing one or more images of an original sequence of images, the sequence of images; or
subtracting, before the determining the image features, a reference image from each image of the sequence of images.
6. The method of claim 1 , further comprising:
determining, based on pixel data from the sequence of images and by application of one or more convolutional layers, the spatial data and the sequential data.
7. The method of claim 1 , further comprising:
applying, to the spatially filtered spatial data and to obtain the sequential data, an attention filter associated with a plurality of probabilities of motion of the one or more objects.
8. The method of claim 1 , wherein the one or more of the images comprise at least one image of the one or more objects.
9. The method of claim 1 , wherein the sequence of images is based on video content received from a security system.
10. The method of claim 1 , further comprising:
indicating, based on determining that the sequence of images indicates motion of one or more objects, a type of object of the one or more objects.
11. An apparatus comprising:
one or more processors; and
memory storing instructions that, when executed by the one or more processors, cause the apparatus to:
receive a sequence of images;
determine image features of the sequence of images based on:
applying a spatial filter that filters spatial data of the sequence of images; and
after applying the spatial filter, applying a sequential filter that filters sequential data of the sequence of images;
determine, based on the image features, that the sequence of images indicates motion of one or more objects; and
output, based on a query associated with the one or more objects, one or more of the images.
12. The apparatus of claim 11 , wherein the spatial filter reduces the spatial data and the sequential filter reduces the sequential data.
13. The apparatus of claim 11 , wherein the spatial filter reduces the spatial data based on at least one of:
reducing a quantity of data values, in each of a plurality of data groups associated with pixel dimensions of the images, based on data values of the data group; or
determining a data value, for each of a plurality of data groups associated with pixel dimensions of the images, based on a maximum of data values of the data group.
14. The apparatus of claim 11 , wherein the sequential filter reduces the sequential data based on at least one of:
reducing a quantity of data values, in each of a plurality of data groups associated with the sequence of images, based on data values of the data group; or
determining a data value, for each of a plurality of data groups associated with the sequence of images, based on a maximum of data values of the data group.
15. The apparatus of claim 11 , wherein the instructions, when executed by the one or more processors, further cause the apparatus to do one or more of:
determining, before determining the image features and by removing one or more images of an original sequence of images, the sequence of images; or
subtracting, before determining the image features, a reference image from each image of the sequence of images.
16. The apparatus of claim 11 , wherein the instructions, when executed by the one or more processors, further cause the apparatus to:
determine, based on pixel data from the sequence of images and by application of one or more convolutional layers, the spatial data and the sequential data.
17. The apparatus of claim 11 , wherein the instructions, when executed by the one or more processors, further cause the apparatus to:
apply, to the spatially filtered spatial data and to obtain the sequential data, an attention filter associated with a plurality of probabilities of motion of the one or more objects.
18. The apparatus of claim 11 , wherein the one or more of the images comprise at least one image of the one or more objects.
19. The apparatus of claim 11 , wherein the sequence of images is based on video content received from a security system.
20. The apparatus of claim 11 , wherein the instructions, when executed by the one or more processors, further cause the apparatus to:
indicate, based on determining that the sequence of images indicates motion of one or more objects, a type of object of the one or more objects.
21. A non-transitory computer-readable medium having stored thereon computer-executable instructions that, when executed, cause:
receiving, by a computing device, a sequence of images;
determining image features of the sequence of images based on:
applying a spatial filter that filters spatial data of the sequence of images; and
after applying the spatial filter, applying a sequential filter that filters sequential data of the sequence of images;
determining, based on the image features, that the sequence of images indicates motion of one or more objects; and
outputting, based on a query associated with the one or more objects, one or more of the images.
22. The non-transitory computer-readable medium of claim 21 , wherein the spatial filter reduces the spatial data and the sequential filter reduces the sequential data.
23. The non-transitory computer-readable medium of claim 21 , wherein the spatial filter reduces the spatial data based on at least one of:
reducing a quantity of data values, in each of a plurality of data groups associated with pixel dimensions of the images, based on data values of the data group; or
determining a data value, for each of a plurality of data groups associated with pixel dimensions of the images, based on a maximum of data values of the data group.
24. The non-transitory computer-readable medium of claim 21 , wherein the sequential filter reduces the sequential data based on at least one of:
reducing a quantity of data values, in each of a plurality of data groups associated with the sequence of images, based on data values of the data group; or
determining a data value, for each of a plurality of data groups associated with the sequence of images, based on a maximum of data values of the data group.
25. The non-transitory computer-readable medium of claim 21 , wherein the computer-executable instructions, when executed, further cause one or more of:
determining, before the determining the image features and by removing one or more images of an original sequence of images, the sequence of images; or
subtracting, before the determining the image features, a reference image from each image of the sequence of images.
26. The non-transitory computer-readable medium of claim 21 , wherein the computer-executable instructions, when executed, further cause:
determining, based on pixel data from the sequence of images and by application of one or more convolutional layers, the spatial data and the sequential data.
27. The non-transitory computer-readable medium of claim 21 , wherein the computer-executable instructions, when executed, further cause:
applying, to the spatially filtered spatial data and to obtain the sequential data, an attention filter associated with a plurality of probabilities of motion of the one or more objects.
28. The non-transitory computer-readable medium of claim 21 , wherein the one or more of the images comprise at least one image of the one or more objects.
29. The non-transitory computer-readable medium of claim 21 , wherein the sequence of images is based on video content received from a security system.
30. The non-transitory computer-readable medium of claim 21 , wherein the computer-executable instructions, when executed, further cause:
indicating, based on determining that the sequence of images indicates motion of one or more objects, a type of object of the one or more objects.