Picture processing method for correcting distorted pictures and apparatus for executing this method

An image analyzer for use with a display to display a picture and an input unit to enter data, said analyzer receiving a first set of picture data representing an uncorrected picture, said analyzer supplying a second set of picture data representing a corrected picture, each of said first and second sets of picture data comprising a plurality of pixels and each pixel of said first and second sets of picture data holding at least one kind of data, said analyzer comprising:

• a picture input unit to receive said first set of picture data;
• a motion vector designating unit to display on said display said uncorrected picture in accordance with said first set of picture data, and to determine motion vectors for said uncorrected picture on the basis of information from said input unit;
• a picture converting unit to generate a vector relating a pixel in said first set of picture data to a corresponding pixel in said second set of picture data using a weighted vector sum of motion vectors, and to designate data in each pixel of said second set of picture data on the basis of data held by a corresponding pixel of said first set of picture data;
• and
• a picture output unit to output said second set of picture data;
• and
• wherein the weight of said motion vectors is proportional to the reciprocal of the square of r, r representing a distance between a pixel of said second set of picture data and an end point of each of said motion vectors.

An image analyzer for use with a display to display a picture and an input unit to enter data, said analyzer receiving a first set of picture data representing an uncorrected picture, said analyzer supplying a second set of picture data representing a corrected picture, each of said first and second sets of picture data comprising a plurality of pixels and each pixel of said first and second sets of picture data holding at least one kind of data, said analyzer comprising:

• a picture input unit to receive said first set of picture data;
• a motion vector designating unit to display on said display said uncorrected picture in accordance with said first set of picture data and to input motion vectors which have been specified by instruction from the input unit for a plurality of starting points in said uncorrected picture and, when no motion vector has been specified by instruction from the input unit for a particular starting point, to generate a motion vector for said particular starting point using a weighted vector sum of at least two motion vectors which have been specified by instruction from the input unit;
• a picture converting unit to identify a pixel in said first set of picture data corresponding to a pixel in said second set of picture data on the basis of said motion vectors, and to designate data in each pixel of said second set of picture data on the basis of data held by a corresponding pixel of said first set of picture data;
• and
• a picture output unit to output said second set of picture data,
• wherein a weight of each motion vector to be subjected to weighting is determined on the basis of a distance between a starting point of this motion vector and said particular starting point, and
• wherein a weight of each of said motion vectors is the reciprocal of the square of r, r representing the distance between the starting point of said motion vector and said particular starting point.

An image analysis process executed by a computer connected to a display to display a picture and an input unit to enter data, said image analysis process receiving a first set of picture data representing an uncorrected picture, said image analysis process supplying a second set of picture data representing a corrected picture, each of said first and second sets of picture data comprising a plurality of pixels and each pixel of said first and second sets of picture data holding at least one kind of data, said process comprising:

• a first step to receive said first set of picture data;
• a second step to display on said display said uncorrected picture in accordance with said first set of picture data;
• a third step to input motion vectors which have been specified by instruction from the input unit for a plurality of starting points in said uncorrected picture and, when no motion vector has been specified by instruction from the input unit for a particular starting point, to generate a motion vector for said particular starting point using a weighted vector sum of at least two motion vectors which have been specified by instruction from the input unit;
• a fourth step to identify a pixel in said first set of picture data corresponding to a pixel in said second set of picture data on the basis of said motion vectors;
• a fifth step to designate data in each pixel of said second set of picture data, the data designated in each pixel being determined on the basis of data held by a corresponding pixel of said first set of picture data;
• and
• a sixth step to output said second set of picture data;
• wherein a weight of each motion vector to be subjected to weighting is determined on the basis of a distance from a starting point of this motion vector and that of said particular starting point, wherein a weight of each of said motion vectors is the reciprocal of the square of r, r representing a distance between a starting point of said motion vector and that of said particular starting point.

A fingerprint analyzer for use with a display to display a picture, an input unit to enter data, and a memory to store reference fingerprint patterns, said analyzer receiving a first set of picture data representing an uncorrected picture of a latent fingerprint, said analyzer supplying a second set of picture data representing a corrected picture of the latent fingerprint, each of said first and second sets of picture data comprising a plurality of pixels and each pixel of said first and second sets of picture data holding at least one kind of data, said analyzer comprising:

• a picture input unit to receive said first set of picture data representing an uncorrected picture of a latent fingerprint;
• a motion vector designating unit to display, on said display, said uncorrected picture of the latent fingerprint in accordance with said first set of picture data, and to determine motion vectors for said uncorrected picture on the basis of information from said input unit;
• a picture converting unit to identify a pixel in said second set of picture data representing a corrected picture of the latent fingerprint on the basis of said motion vectors, and to designate data in each pixel of said second set of picture data, the data to be designated in each pixel being determined on the basis of data held by the corresponding pixel of said first set of picture data;
• and
• a picture matching unit to match said second set of picture data representing the corrected picture of the latent fingerprint with a reference fingerprint pattern stored in said memory,
• wherein said picture converting unit converts said first picture into said second picture using a vector between a pixel in said second set of picture data and a corresponding pixel in said first set of picture data obtained by a weighted vector sum of motion vectors, and
• wherein the weight of each of said motion vectors is proportional to the reciprocal of the square of r, r representing the distance between the pixel of said second set of picture data and the end point of each of said motion vectors.

A fingerprint analyzer for use with a display to display a picture, an input unit to enter data, and a memory to store reference fingerprint patterns, said analyzer receiving a first set of picture data representing an uncorrected picture of a latent fingerprint, said analyzer supplying a second set of picture data representing a corrected picture of the latent fingerprint, each of said first and second sets of picture data comprising a plurality of pixels and each pixel of said first and second sets of picture data holding at least one kind of data, said analyzer comprising:

• a picture input unit to receive said first set of picture data representing an uncorrected picture of a latent fingerprint;
• a motion vector designating unit to display, on said display, said uncorrected picture of the latent fingerprint in accordance with said first set of picture data, and to determine motion vectors for said uncorrected picture on the basis of information from said input unit;
• a picture converting unit to identify a pixel in said second set of picture data representing a corrected picture of the latent fingerprint on the basis of said motion vectors, and to designate data in each pixel of said second set of picture data, the data to be designated in each pixel being determined on the basis of data held by the corresponding pixel of said first set of picture data;
• and
• a picture matching unit to match said second set of picture data representing the corrected picture of the latent fingerprint with a reference fingerprint pattern stored in said memory,
• wherein the motion vector designating unit sets a plurality of starting points in said uncorrected picture, and determines motion vectors each corresponding to one of said plurality of starting points,
• wherein said motion vector designating unit, if said plurality of starting points include an unidentified starting point for which no motion vector has been determined, calculates the motion vector for said unidentified starting point on the basis of determined motion vectors,
• wherein
• said motion vector designating unit obtains the motion vector for said unidentified starting point by weight averaging of the determined motion vectors, and
• the weight of each motion vector to be subjected to weighted averaging is determined on the basis of the distance from the starting point of this motion vector and that of said unidentified starting point, and
• wherein the weight of each of said motion vectors is the reciprocal of the square of r, r representing the distance between the starting point of said motion vector and that of said unidentified starting point.

A fingerprint matching process executed by a computer connected to a display to display a picture, an input unit to enter data, and a memory to store reference fingerprint patterns, said process receiving a first set of picture data representing an uncorrected picture of a latent fingerprint, said process supplying a second set of picture data representing a corrected picture of the latent fingerprint, each of said first and second sets of picture data comprising a plurality of pixels and each pixel of said first and second sets of picture data holding at least one kind of data, said process comprising:

• a first step to receive said first set of picture data;
• a second step to display on said display said uncorrected picture of a latent fingerprint in accordance with said first set of picture data;
• a third step to determine motion vectors for said uncorrected picture of the latent fingerprint on the basis of information from said input unit;
• a fourth step to identify a pixel in said first set of picture data corresponding to a pixel in said second set of picture data on the basis of said motion vectors;
• a fifth step to designate data in each pixel of said second set of picture data, the data to be designated in each pixel being determined on the basis of data held by the corresponding pixel of said first set of picture data;
• and
• a sixth step to match said second set of picture data representing a corrected picture of the latent fingerprint with a reference fingerprint pattern stored in said memory,
• wherein a vector between a pixel in said second set of picture data and a corresponding pixel in said first set of picture data is obtained at said fourth step by a weighted vector sum of motion vectors, and
• wherein the weight of each of said motion vectors is proportional to the reciprocal of the square of r, r representing the distance between the pixel of said second set of picture data and the end point of each of said motion vectors.

A fingerprint matching process executed by a computer connected to a display to display a picture, an input unit to enter data, and a memory to store reference fingerprint patterns, said process receiving a first set of picture data representing an uncorrected picture of a latent fingerprint, said process supplying a second set of picture data representing a corrected picture of the latent fingerprint, each of said first and second sets of picture data comprising plurality of pixels and each pixel of said first and second sets of picture data holding at least one kind of data said process comprising:

• a first step to receive said first set of picture data;
• a second step to display on said display said uncorrected picture of a latent fingerprint in accordance with said first set of picture data;
• a third step to determine motion vectors for said uncorrected picture of the latent fingerprint on the basis of information from said input unit;
• a fourth step to identify a pixel in said first set of picture data corresponding to a pixel in said second set of picture data on the basis of said motion vectors;
• a fifth step to designate data in each pixel of said second set of picture data, the data to be designated in each pixel being determined on the basis of data held by the corresponding pixel of said first set of picture data;
• and
• a sixth step to match said second set of picture data representing a corrected picture of the latent fingerprint with a reference fingerprint pattern stored in said memory,
• wherein said third step includes
• a step to set a plurality of starting points in said uncorrected picture, and
• a step to determine motion vectors, each motion vector corresponding to one of said plurality of starting points, on the basis of instruction from said input unit,
• wherein, if said plurality of starting points include any unidentified starting point for which no motion vector has been determined, the motion vector for said unidentified starting point is calculated on the basis of determined motion vectors,
• wherein
• the motion vector for said unidentified starting point is obtained by weighted averaging of the determined motion vectors,
• the weight of each motion vector to be subjected to weighted averaging is determined on the basis of the distance from the starting point of this motion vector and that of said unidentified starting point, and
• wherein the weight of each of said motion vectors is the reciprocal of the square of r, r representing the distance between the starting point of said motion vector and that of said unidentified starting point.