Adaptive video compression

1. A computer-implemented method, comprising:

determining one or more feature values of video data;

analyzing the one or more feature values using a model trained to classify the video data into one of a plurality of classes associated with a plurality of respective encoding schemes, wherein the plurality of classes correspond to a plurality of clusters of sample video data formed during unsupervised training of the model the encoding scheme of a class determined based at least in part on feature values of sample video data in the cluster that corresponds to the class;

classifying the video data into a class selected from the plurality of classes using the model, the class associated with an encoding scheme; and

encoding the video data using the encoding scheme.

2. The method of claim 1 , further comprising:

down-sampling the video data to reduce the sampling rate of the video data.

3. The method of claim 1 , further comprising:

processing the video data using a preliminary encoding process to compress the video data and determine at least one of the one or more feature values.

4. The method of claim 1 , wherein the one or more feature values includes one or more video complexity values.

5. The method of claim 4 , wherein the one or more video complexity values includes a degree of complexity within a frame of the video data and a degree of complexity between multiple frames of the video data.

6. The method of claim 1 , further comprising:

obtaining the plurality of sample video data;

determining respective feature vectors for the plurality of sample video data;

determining a plurality of clusters of features vectors;

defining the plurality of classes based on the plurality of clusters; and

associating the plurality of classes with respective encoding schemes.

7. The method of claim 6 , further comprising:

determining respective model feature vectors for the plurality of classes, wherein the model includes at least the plurality of classes and respective model feature vectors;

generating a feature vector for the received video data;

comparing the feature vector to the model feature vectors to determine the class of the received video data.

8. The method of claim 1 , further comprising:

receiving a video file;

segmenting the video file into a plurality of segments of video data;

determining one or more feature values for individual segments of the plurality of segments;

classifying the individual segments into respective classes using the model, the classes associated with respective encoding schemes;

encoding the individual segments using the respective encoding schemes to generate a plurality of compressed segments; and

assembling the plurality of compressed segments into a continuous data stream.

9. The method of claim 8 , further comprising:

encoding at least two of the plurality of segments using different encoding schemes.

10. The method of claim 8 , further comprising:

segmenting the video file in the plurality of segments based at least in part on one or more encoder-determined compression characteristics of the video file.

11. The method of claim 1 , further comprising:

receiving the video data from a third party provider; and

transmitting the encoded video data to the third party provider or a user device associated with the third party provider.

12. A system, comprising:

at least one computing device processor; and

a memory device including instructions that, when executed by the at least one computing device processor, cause the system to:

receive video data;

determine one or more feature values of the video data;

analyze the one or more feature values using a model trained to classify the video data into one of a plurality of classes associated with a plurality of respective encoding schemes, wherein the plurality of classes correspond to a plurality of clusters of sample video data formed during unsupervised training of the model the encoding scheme of a class determined based at least in part on feature values of sample video data in the cluster that corresponds to the class;

classify the video data into a class selected from the plurality of classes using the model, the class associated with an encoding scheme; and

encode the video data using the encoding scheme.

13. The system of claim 12 , wherein the one or more feature values includes one or more video complexity values, including a degree of complexity within a frame of the video data and a degree of complexity between multiple frames of the video data.

14. The system of claim 12 , wherein the instructions when executed further cause the system to:

obtain the plurality of training video data;

determine respective feature vectors for the plurality of training video data;

determine a plurality of clusters of features vectors;

define the plurality of classes based on the plurality of clusters; and

associate the plurality of classes with respective encoding.

15. The system of claim 14 , wherein the instructions when executed further cause the system to:

determine respective model feature vectors for the plurality of classes, wherein the model includes at least the plurality of classes and respective model feature vectors;

generate a feature vector for the received video data;

compare the feature vector to the model feature vectors to determine the class of the received video data.

16. The system of claim 12 , wherein the instructions when executed further cause the system to:

receive a video file;

segment the video file into a plurality of segments of video data;

determine one or more feature values for individual segments of the plurality of segments;

classify the individual segments into respective classes using the model, the classes associated with respective encoding schemes;

encode the individual segments using the respective encoding schemes to generate a plurality of compressed segments; and

combine the plurality of compressed segments into a continuous data stream.

17. A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to:

receive video data;

determine one or more feature values of the video data;

analyze the one or more feature values using a model trained to classify the video data into one of a plurality of classes associated with a plurality of respective encoding schemes, wherein the plurality of classes correspond to a plurality of clusters of sample video data formed during unsupervised training of the model the encoding scheme of a class determined based at least in part on feature values of sample video data in the cluster that corresponds to the class;

classify the video data into a class selected from the plurality of classes using the model, the class associated with an encoding scheme; and

encode the video data using the encoding scheme.

18. The non-transitory computer-readable storage medium of claim 17 , wherein the one or more feature values includes one or more video complexity values, including a degree of complexity within a frame of the video data and a degree of complexity between multiple frames of the video data.

19. The non-transitory computer-readable storage medium of claim 17 , wherein the instructions when executed further cause the system to:

obtain the plurality of training video data;

determine respective feature vectors for the plurality of training video data;

determine a plurality of clusters of features vectors;

define the plurality of classes based on the plurality of clusters; and

associate the plurality of classes with respective encoding.

20. The non-transitory computer-readable storage medium of claim 19 , wherein the instructions when executed further cause the system to:

determine respective model feature vectors for the plurality of classes, wherein the model includes at least the plurality of classes and respective model feature vectors;

generate a feature vector for the received video data;

compare the feature vector to the model feature vectors to determine the class of the received video data.