Program, information processing apparatus and information processing method to change location information of slideshow

A computer program product having a non-transitory computer readable medium including programmed instructions stored thereon, wherein the instructions, when executed by a computer, cause the computer to function as:

• a memory control unit that reads a first stride parameter used for controlling an output resolution and a first dilation parameter used for controlling an input resolution from a memory device;
• a transformation unit that transforms the first stride parameter to a second stride parameter and transforms the first dilation parameter to a second dilation parameter by using a transformation parameter;
• a first convolutional neural network (CNN) unit that executes first CNN processing of a feature vector by using at least the second stride parameter;
• and
• a second CNN unit that executes second CNN processing with an output vector of the first CNN unit as an input by using at least the second dilation parameter,
• wherein:
• the feature vector represents a feature of an input audio, and
• the instructions cause the computer to further function as:
• an extraction unit that generates the feature vector by extracting the feature from the input audio, and
• a recognition unit that recognizes the audio by using an output vector of the second CNN unit.

An information processing method implemented by an information processing apparatus, the method comprising:

• reading a first stride parameter used for controlling an output resolution and a first dilation parameter used for controlling an input resolution from a memory device;
• transforming the first stride parameter to a second stride parameter and transforming the first dilation parameter to a second dilation parameter by using a transformation parameter;
• executing first convolutional neural network (CNN) processing of a feature vector by using at least the second stride parameter;
• and
• executing second CNN processing with an output vector of the first CNN processing as an input by using at least the second dilation parameter,
• wherein:
• the feature vector represents a feature of an input audio, and
• the method further comprises:
• generating the feature vector by extracting the feature from the input audio, and
• recognizing the audio by using an output vector of the second CNN processing.

An information processing method implemented by an information processing apparatus, the method comprising:

• reading a first stride parameter used for controlling an output resolution and a first dilation parameter used for controlling an input resolution from a memory device;
• transforming the first stride parameter to a second stride parameter and transforming the first dilation parameter to a second dilation parameter by using a transformation parameter;
• storing at least the second stride parameter in a second information processing apparatus as a parameter used in first convolutional neural network (CNN) processing of a feature vector;
• and
• storing at least the second dilation parameter in the second information processing apparatus as a parameter used in second CNN processing with an output vector of the first CNN processing as an input,
• wherein:
• the feature vector represents a feature of an input audio, and
• the method further comprises:
• generating the feature vector by extracting the feature from the input audio, and
• recognizing the audio by using an output vector of the second CNN processing.