Data composite for efficient memory transfer in a behavioral recognition system
View Patent ↗Techniques are disclosed for dynamic memory allocation in a behavioral recognition system. According to one embodiment of the disclosure, one or more variable-sized chunks of memory is allocated from a device memory for a memory pool. An application allocates at least one of the chunks of memory from the memory pool for processing a plurality of input data streams in real-time. A request to allocate memory from the memory pool for input data is received. Upon determining that one of the chunks is available in the memory pool to store the input data, the chunk is allocated from the memory pool in response to the request.
1. A computer-implemented method comprising:
receiving, from at least one data provider, a plurality of video feeds, at least one video feed of the plurality of video feeds having at least one of a resolution and a frame rate that is different relative to other video feeds of the plurality of video feeds;
allocating, from a device memory, at least one chunk of memory from one or more variable-sized chunks of available memory of a memory pool;
packaging the plurality of video feeds as a composite of the plurality of video feeds into the at least one allocated chunk of memory of the memory pool; and
transferring the packaged composite of the plurality of video feeds to a machine-learning engine, the machine-learning engine including a neuro-linguistic model configured to determine anomalous behavior in the plurality of video feeds.
2. The method of claim 1 , wherein allocating at least one chunk of memory from one or more variable-sized chunks of available memory includes determining that at least one chunk of memory of the memory pool is not available.
3. The method of claim 1 , further comprising:
returning the at least one allocated chunk of memory to the memory pool after the packaged composite of the plurality of video feeds has been transferred.
4. The method of claim 1 , wherein the at least one allocated chunk of memory is associated with a time decay constant.
5. The method of claim 4 , further comprising:
releasing the at least one allocated chunk of memory after the at least one allocated chunk is unused after a specified amount of time based on the time decay constant.
6. The method of claim 1 , wherein a device associated with the device memory is a graphics processing unit (GPU).
7. The method of claim 1 , wherein the at least one allocated chunk of memory is allocated based on a total device memory size.
8. A non-transitory computer-readable storage medium comprising: computer-executable instructions that, when executed on at least one processor, cause the at least one processor to:
receive, from at least one data provider, a plurality of video feeds, at least one video feed of the plurality of video feeds having at least one of a resolution and a frame rate that is different relative to other video feeds of the plurality of video feeds;
allocate, from a device memory, at least one chunk of memory from one or more variable-sized chunks of available memory of a memory pool;
package the plurality of video feeds as a composite of the plurality of video feeds into the at least one allocated chunk of memory of the memory pool; and
transfer the packaged composite of the plurality of video feeds to a machine-learning engine, the machine-learning engine including a neuro-linguistic model configured to determine anomalous behavior in the plurality of video feeds.
9. The non-transitory computer-readable storage medium of claim 8 , wherein the computer-executable instructions to allocate the at least one chunk of memory include instructions to:
determine that at least one chunk of memory of the memory pool is not available for allocation.
10. The non-transitory computer-readable storage medium of claim 8 , wherein the computer-executable instructions further comprise instructions that cause the at least one processor to:
return the at least one allocated chunk of memory to the memory pool.
11. The non-transitory computer-readable storage medium of claim 8 , wherein the at least one chunk of allocated memory is associated with a time decay constant.
12. The non-transitory computer-readable storage medium of claim 11 , wherein the computer-executable instructions further comprise instructions that cause the at least one processor to:
release the at least one allocated chunk of memory if the at least one allocated chunk of memory is unused after a specified amount of time based on the time decay constant.
13. The non-transitory computer-readable storage medium of claim 8 , wherein the device memory is associated with a graphics processing unit (GPU).
14. The non-transitory computer-readable storage medium of claim 8 , wherein the at least one allocated chunk of memory is allocated based on a total device memory size.
15. A system, comprising:
a processor; and
a memory in communication with the processor and storing code, which, when executed on the processor, causes the processor to:
receive a plurality of video feeds, at least one video feed of the plurality of video feeds having at least one of a resolution and a frame rate that is different relative to other video feeds of the plurality of video feeds,
allocate at least one chunk of memory from one or more variable-sized chunks of available memory of a memory pool,
package the plurality of video feeds as a composite of the plurality of video feeds into the at least one allocated chunk of memory, and
transfer a copy of the packaged composite of the plurality of video feeds to a machine-learning engine including a neuro-linguistic model configured to determine anomalous behavior in the plurality of video feeds.
16. The system of claim 15 , wherein the code to allocate the at least one chunk of memory includes code to:
determine if at least one chunk of memory of the memory pool is not available for allocation.
17. The system of claim 15 , wherein the code, when executed, further causes the processor to:
return the at least one allocated chunk of memory to the memory pool.
18. The system of claim 15 , wherein the at least one chunk of allocated memory is associated with a time decay constant.
19. The system of claim 18 , wherein the code, when executed, further causes the processor to:
release the at least one allocated chunk of memory if the at least one allocated chunk of memory is unused after a specified amount of time based on the time decay constant.
20. The system of claim 15 , further comprising a graphics processing unit (GPU).