Methods and systems for counting people

A system comprising:

• a memory;
• and
• a processor coupled to the memory and configured to perform a method comprising:
• pre-training an analytics system, wherein
• the pre-training comprises
• generating an augmented data set, wherein
• the generating the augmented data set comprises injecting one or more modifications into a ground truth data set, and
• the one or more modifications are selected by a domain expert,
• generating a semantic fingerprint that comprises a probability distribution, and
• utilizing the augmented data set and the semantic fingerprint to perform an unsupervised generative pre-training process and a supervised discriminative training fine-tuning process;
• receiving input data at the processor, wherein
• the input data comprises a representation of an environment at a first point in time,
• the representation of the environment is based, at least in part, on data captured by one or more sensors configured to capture visual data,
• the representation of the environment comprises a plurality of representations of persons, and
• the system comprises an artificial neural network;
• and
• subsequent to the pre-training, using the analytics system to perform operations comprising:
• pre-processing the input data, wherein
• the pre-processing comprises whitening the input data, wherein
• whitening the input data comprises transforming the input data to uncorrelated values,
• identifying, using the processor, a representation of a person of the plurality of representations of persons, wherein
• identifying the representation of the person comprises
• generating a feature vector using a generative process, wherein
• the feature vector comprises a sparse distributed representation of the input data, and
• applying a discriminative process to the feature vector to identify the representation of the person,
• generating a count of a number of persons identified as being present within the input data, wherein
• the count is generated, at least in part, by using the representation of the person, and
• updating a value, based, at least in part, on the count, wherein
• the value indicates the number of persons identified as being present within the input data.

A non-transitory computer readable storage medium comprising program instructions executable to perform a method comprising:

• pre-training an analytics system, wherein
• the pre-training comprises
• generating an augmented data set, wherein
• the generating the augmented data set comprises injecting one or more modifications into a ground truth data set, and
• the one or more modifications are selected by a domain expert,
• generating a semantic fingerprint that comprises a probability distribution, and
• utilizing the augmented data set and the semantic fingerprint to perform an unsupervised generative pre-training process and a supervised discriminative fine-tuning process;
• receiving input data at a processor of the analytics system, wherein
• the input data comprises a representation of an environment at a first point in time,
• the representation of the environment is based, at least in part, on data captured by one or more sensors configured to capture visual data,
• the representation of the environment comprises a plurality of representations of persons, and
• the analytics system comprises an artificial neural network;
• and
• subsequent to the pre-training, using the analytics system to perform operations comprising:
• pre-processing the input data, wherein
• the pre-processing comprises whitening the input data, wherein
• whitening the input data comprises transforming the input data to uncorrelated values,
• identifying, using the processor, a representation of a person of the plurality of representations of persons, wherein
• identifying the representation of the person comprises
• generating a feature vector using a generative process, wherein
• the feature vector comprises a sparse distributed representation of the input data, and
• applying a discriminative process to the feature vector to identify the representation of the person,
• generating a count of a number of persons identified as being present within the input data, wherein
• the count is generated, at least in part, by using the representation of the person, and
• updating a value, based, at least in part, on the count, wherein
• the value indicates the number of persons identified as being present within the input data.