Abstract: A smart roller comprises: an exterior annular cylinder portion comprising an elastomeric material and having an exterior cylindrical surface; a sensor array imbedded in a volume of the exterior annular cylinder portion, the sensor array extending in an axial direction and in a circumferential direction of the exterior annular cylinder portion, the array comprising a plurality of independently sampleable sensor elements, each sensor element located for measurement at a corresponding axial and circumferential sensor location; a rigid interior portion, at least a portion of the rigid interior section disposed in a bore of the exterior annular cylinder portion, the rigid interior portion connected to the exterior annular cylinder portion for unitary rotational movement therewith; and readout electronics operably connected to the sensor array and configurable to independently sample sensor output from each of the sensor elements.

Abstract: A method and system for codec of visual feature data is provided. First protocol format visual feature data generated by an intelligent front end and a certificate identity used to uniquely identify a corresponding first protocol format are received. The first protocol format visual feature data is converted to a same type of second protocol format visual feature data according to the certificate identity. The second protocol format visual feature data is received and parsed according to the second protocol format to obtain original visual feature data that are generated by the intelligent front end and contained in the first protocol format visual feature data.

Abstract: Disclosed are a compression method and system for the frequent transmission of a deep neural network. The deep neural network compression is extended to the field of transmission, and the potential redundancy among deep neural network models is utilized for compressing, so that the overhead of the deep neural network under frequency transmission is reduced. The advantages of the present invention are that: in the present invention, the redundancy among multiple models of the deep neural network on the frequency transmission is combined, knowledge information among deep neural networks is utilized for compressing, and the size and the bandwidth of the required transmission are reduced. The deep neural network can be better transmitted under the same bandwidth limitation; meanwhile, the deep neural network is allowed to be performed targeted compression at a front end, rather than being restored partially after being performed targeted compression.

Abstract: A method and system for codec of visual feature data is provided. First protocol format visual feature data generated by an intelligent front end and a certificate identity used to uniquely identify a corresponding first protocol format are received. The first protocol format visual feature data is converted to a same type of second protocol format visual feature data according to the certificate identity. The second protocol format visual feature data is received and parsed according to the second protocol format to obtain original visual feature data that are generated by the intelligent front end and contained in the first protocol format visual feature data.