Methods and systems for detection of chroma dropouts in video

A method for detecting one or more chroma dropout errors in a current field, the current field associated with a video frame, the method comprising:

• grouping a first plurality of pixels associated with the current field into a first set of pixel pairs;
• calculating a first co-occurrence matrix corresponding to a second set of pixel pairs associated with a previous field, wherein each pixel pair of the second set of pixel pairs corresponds in location to each pixel pair of the first set of pixel pairs, wherein the previous field is a field processed previously having a polarity similar to that of the current field, wherein the second set pixel pairs includes a second plurality of pixels;
• calculating a second co-occurrence matrix corresponding to a third set of pixel pairs associated with a subsequent field, wherein each pixel pair of the third set of pixel pairs corresponds in location to each pixel pair of the first set of pixel pairs, wherein the subsequent field is a field processed immediately after processing of the current field having a polarity similar to that of the current field, wherein the third set pixel pairs includes a third plurality of pixels;
• selecting a first pixel pair associated with the first set of pixel pairs, wherein the first pixel pair includes first and second pixels of the first plurality of pixels, wherein the first and second pixels include first and second chroma values, respectively, and wherein the first and second pixels include first and second luma values, respectively;
• selecting a first set of entries of the first co-occurrence matrix based on the first and second chroma values;
• selecting a second set of entries of the second co-occurrence matrix based on the first and second chroma values;
• determining a presence of the first pixel pair in the first and second co-occurrence matrices by analysing the first and second set of entries of the first and second co-occurrence matrices, respectively;
• calculating a first absolute difference between the first and second chroma values;
• comparing the first absolute difference with a first predetermined threshold;
• populating first and second tables that include one or more chroma values associated with the second and third pluralities of pixels, respectively, and a count of pixels corresponding to each of the one or more chroma values associated with the second and third pluralities of pixels, respectively;
• sorting each of the first and second tables in an ascending order based on the one or more chroma values associated with the second and third pluralities of pixels, respectively;
• selecting a first minimum chroma value from the first table that is a smallest chroma value having a corresponding count of pixels at least one of greater than and equal to a predetermined value;
• selecting a first maximum chroma value from the first table that is a largest chroma value having a corresponding count of pixels at least one of greater than and equal to the predetermined value;
• selecting a second minimum chroma value from the second table that is a smallest chroma value having a corresponding count of pixels at least one of greater than and equal to the predetermined value;
• selecting a second maximum chroma value from the second table that is a largest chroma value having a corresponding count of pixels at least one of greater than and equal to the predetermined value;
• calculating a minimum chroma value that is smaller of the first and second minimum chroma values;
• calculating a maximum chroma value that is larger of the first and second maximum chroma values;
• comparing the first and second chroma values with the minimum and maximum chroma values;
• calculating a second absolute difference between the first and second luma values;
• comparing the second absolute difference with a second predetermined threshold;
• marking a first template block of a predefined size and having a centre within a first predetermined distance from first and second pixels;
• obtaining a first motion compensated template block in the previous field corresponding to the first template block;
• obtaining a second motion compensated template block in the subsequent field corresponding to the first template block;
• calculating a first matching metric corresponding to a first plurality of luma values associated with corresponding pixels of the first template block and the first motion compensated template block;
• calculating a second matching metric corresponding to a second plurality of luma values associated with corresponding pixels of the first template block and the second motion compensated template block;
• marking the first pixel pair as erroneous when the first and second matching metrics are less than a third predetermined threshold;
• increasing an error count corresponding to the current field when the first and second matching metrics are less than the third predetermined threshold;
• and
• marking the current field as erroneous when the error count is greater than a fourth predetermined threshold.

A system for detecting one or more chroma dropout errors in a current field, the current field associated with a video frame, the system comprising:

• a. a pixel-pair creation module for grouping a first plurality of pixels associated with the current field into a first set of pixel pairs;
• b. a co-occurrence matrix calculation module for performing steps comprising:
• calculating a first co-occurrence matrix corresponding to a second set of pixel pairs associated with a previous field, wherein each pixel pair of the second set of pixel pairs corresponds in location to each pixel pair of the first set of pixel pairs, wherein the previous field is a field processed previously having a polarity similar to that of the current field, wherein the second set pixel pairs includes a second plurality of pixels;
• and
• calculating a second co-occurrence matrix corresponding to a third set of pixel pairs associated with a subsequent field, wherein each pixel pair of the third set of pixel pairs corresponds in location to each pixel pair of the first set of pixel pairs, wherein the subsequent field is a field processed immediately after processing of the current field having a polarity similar to that of the current field, wherein the third set pixel pairs includes a third plurality of pixels;
• and
• c. a chroma dropout error identification module, comprising:
• i. Pixel pair search module for performing steps comprising:
• selecting a first pixel pair associated with the first set of pixel pairs, wherein the first pixel pair includes first and second pixels of the first plurality of pixels, wherein the first and second pixels include first and second chroma values, respectively, and wherein the first and second pixels include first and second luma values, respectively;
• selecting a first set of entries of the first co-occurrence matrix based on the first and second chroma values;
• determining a presence of the first pixel pair in the first co-occurrence matrix by analysing the first set of entries of the first co-occurrence matrix;
• ii. Effective minimum and maximum computation module for performing steps comprising:
• populating a first table that includes one or more chroma values associated with the second plurality of pixels and a count of pixels corresponding to each of the one or more chroma values;
• sorting the first table in an ascending order based on the one or more chroma values;
• selecting a first minimum chroma value from the first table that is a smallest chroma value having a corresponding count of pixels at least one of greater than and equal to a predetermined value;
• selecting a first maximum chroma value from the first table that is a largest chroma value having a corresponding count of pixels at least one of greater than and equal to the predetermined value;
• and
• comparing the first and second chroma values with the first minimum and maximum chroma values;
• iii. Gradient computation and comparison module for performing steps comprising:
• calculating a second absolute difference between the first and second luma values;
• and
• comparing the second absolute difference with a second predetermined threshold;
• calculating a first absolute difference between the first and second chroma values;
• and
• comparing the first absolute difference with a first predetermined threshold;
• iv. Template block identification module for marking a first template block of a predefined size and having a centre within a first predetermined distance from first and second pixels;
• v. Motion estimation module for obtaining a first motion compensated template block in the previous field corresponding to the first template block;
• and
• vi. Matching metric computation module for performing the steps comprising:
• calculating a first matching metric corresponding to a first plurality of luma values associated with corresponding pixels of the first template block and the first motion compensated template block;
• marking the first pixel pair as erroneous when the first matching metric is less than a third predetermined threshold;
• increasing an error count corresponding to the current field when the first matching metric is less than the third predetermined threshold;
• and
• marking the current field as erroneous when the error count is greater than a fourth predetermined threshold.

A computer program product comprising computer-executable instructions embodied in a non-transitory computer-readable medium for use in connection with a processor-containing system, for executing steps comprising:

• grouping a first plurality of pixels associated with the current field into a first set of pixel pairs;
• calculating a first co-occurrence matrix corresponding to a second set of pixel pairs associated with a previous field, wherein each pixel pair of the second set of pixel pairs corresponds in location to each pixel pair of the first set of pixel pairs, wherein the previous field is a field processed previously having a polarity similar to that of the current field, wherein the second set pixel pairs includes a second plurality of pixels;
• calculating a second co-occurrence matrix corresponding to a third set of pixel pairs associated with a subsequent field, wherein each pixel pair of the third set of pixel pairs corresponds in location to each pixel pair of the first set of pixel pairs, wherein the subsequent field is a field processed immediately after processing of the current field having a polarity similar to that of the current field, wherein the third set pixel pairs includes a third plurality of pixels;
• selecting a first pixel pair associated with the first set of pixel pairs, wherein the first pixel pair includes first and second pixels of the first plurality of pixels, wherein the first and second pixels include first and second chroma values, respectively, and wherein the first and second pixels include first and second luma values, respectively;
• selecting a first set of entries of the first co-occurrence matrix based on the first and second chroma values;
• selecting a second set of entries of the second co-occurrence matrix based on the first and second chroma values;
• determining a presence of the first pixel pair in the first and second co-occurrence matrices by analyzing the first and second set of entries of the first and second co-occurrence matrices, respectively;
• calculating a first absolute difference between the first and second chroma values;
• comparing the first absolute difference with a first predetermined threshold;
• calculating minimum and maximum chroma values corresponding to the second and third plurality of pixels associated with the previous and subsequent fields, respectively;
• comparing the first and second chroma values with the minimum and maximum chroma values;
• calculating a second absolute difference between the first and second luma values;
• comparing the second absolute difference with a second predetermined threshold;
• marking a first template block having first and second pixels as diagonal vertices;
• obtaining a first motion compensated template block in the previous field corresponding to the first template block;
• obtaining a second motion compensated template block in the subsequent field corresponding to the first template block;
• calculating a first matching metric corresponding to a first plurality of luma values associated with corresponding pixels of the first template block and the first motion compensated template block;
• calculating a second matching metric corresponding to a second plurality of luma values associated with corresponding pixels of the first template block and the second motion compensated template block;
• marking the first pixel pair as erroneous when the first and second matching metrics are less than a third predetermined threshold;
• increasing an error count corresponding to the current field when the first and second matching metrics are less than the third predetermined threshold;
• and
• marking the current field as erroneous when the error count is greater than a fourth predetermined threshold.