Optimizing the displaying of HDR images
1. To obtain in a pragmatic manner better watchable HDR images, we propose several variants of an image processing apparatus for processing an input image (Im_Comm) of an input video to obtain an output image of an output video, wherein the input image has pixels which have input luminances (Ln_in_pix) which fall within a first luminance dynamic range (DR_ 1 ), which first luminance dynamic range has a first maximum luminance (PL_V_HDR), wherein the output image has pixels which have output luminances which can be calculated from the input luminances and which fall within a second luminance dynamic range, which second luminance dynamic range has a second maximum luminance (PL_V_MDR), wherein the apparatus comprises a video data input ( 227 ) arranged to receive the input image and a reference luminance mapping function (F_L) which is encoded as metadata associated with the input image, wherein the reference luminance mapping function specifies a relationship between luminances of a first reference image and luminances of a second reference image, wherein the first reference image is the input image, wherein the second reference image has a second reference maximum luminance (PL_V_SDR); wherein the apparatus comprises a display adaptation unit ( 209 ) arranged to determine an adapted luminance mapping function (FL_DA) which is based on the reference luminance mapping function (F_L), wherein the display adaptation unit uses a pre-fixed display adaptation algorithm, which specifies for each point on a diagonal, in a coordinate system of the input luminances normalized to a maximum being 1 and the output luminances normalized to a maximum being 1, a respective metric along a line segment oriented in a pre-fixed direction, the line segment starting on the diagonal, wherein the respective metric at each position is normalized by giving the point of intersection of the line segment and the locus of the luminance function (F_L) in the coordinate system the value 1, wherein the determining locates a position on each respective metric corresponding to the second maximum luminance (PL_V_MDR), wherein the collection of positions on the respective metrics is output as the adapted luminance mapping function (FL_DA); characterized in that the apparatus comprises a processing comprises a boost determination circuit ( 700 ) arranged to calculate a booster strength value (BO) comprising: —a histogram calculation unit ( 721 ) arranged to determine a histogram (hist) of intermediate luminances obtained by applying the adapted luminance mapping function (FL_DA) to the input luminances, —an average calculation circuit ( 705 ) arranged to calculate an average brightness measure (AB) on the basis of the histogram, —a first converter unit ( 706 ) arranged to calculate a first strength value (PosB) from the average brightness measure (AB), —a second measuring unit ( 711 ) arranged to calculate a weighted sum (BDE) of pixel counts in at least two configurable upper bins of the histogram, —a second conversion unit ( 712 ) arranged to calculate a second strength value (NegB) from the weighted sum (BDE), —wherein the boost determination circuit determines the booster strength value (BO) on the basis of a value (DV) which is equal to the first strength value (PosB) minus the second strength value (NegB); wherein the apparatus is further arranged to: use the display adaptation unit to calculate an adjusted adapted luminance mapping function (F_ALT_B), wherein the adjusted adapted luminance mapping function is calculated by using the display adaptation algorithm in a setting which outputs the locus of positions equal to the booster strength value on the metric as the adjusted adapted luminance mapping function (F_ALT_B); and the apparatus comprising a color transformer ( 208 ) arranged to apply the adjusted adapted luminance mapping function (F_ALT_B) to the input luminances to obtain the output luminances.
1 . A method comprising:
obtaining an input image and metadata,
wherein the input image has pixels which have input luminances,
wherein the input luminances fall within an input luminance dynamic range,
wherein the input luminance dynamic range has an input maximum luminance,
wherein the metadata comprises a reference luminance mapping function,
wherein the reference luminance mapping function specifies a relationship between luminances of the input image and reference luminances,
wherein the reference luminances lie in a reference dynamic range,
wherein the reference dynamic range has a reference maximum luminance;
determining an adapted luminance mapping function,
wherein the adapted luminance mapping function is based on the reference luminance mapping function;
wherein the adapted luminance mapping function is calculated by:
specifying for each point on a diagonal, a metric along a line segment oriented in a direction,
wherein the diagonal is set in coordinate system of the input luminances normalized to a maximum of 1 and output luminances normalized to a maximum of 1,
wherein the line segment has a first endpoint on the diagonal and has a second endpoint at intersection of the line segment and the reference luminance mapping function; and
locating a position on each metric corresponding to an output maximum luminance;
determining a histogram of intermediate luminances by applying the adapted luminance mapping function to the input luminances;
calculating an average brightness measure based on the histogram,
calculating a first strength from the average brightness measure;
calculating a weighted sum of pixel counts in at least two brightest bins of the histogram;
calculating a second strength value from the weighted sum;
determining a booster strength based on a difference, wherein the difference is the first strength minus the second strength;
calculating an adjusted adapted luminance mapping function as a locus of positions of the booster strength on the metric; and
applying the adjusted adapted luminance mapping function on the input luminances so as to obtain output luminances.
2 . The method as claimed in claim 1 , wherein one of the at least two brightest bins of the histogram comprises the 1% brightest pixels in the histogram.
3 . The method as claimed in claim 1 , wherein the booster strength is determined as a multiplier multiplied by a relative boost.
4 . The method as claimed in claim 1 , wherein the calculating of the booster strength is based on temporal filtering so as to limit a change with a previously determined booster strength.
5 . The method as claimed in claim 4 ,
wherein temporal filtering is asymmetric,
wherein the temporal filtering is rapidly changing to lesser booster strength,
wherein the temporal filtering is slower changing to higher booster strength.
6 . An image processing apparatus comprising:
a processor circuit and a memory circuit, wherein the memory circuit is arranged to store at least one instructions for the processor circuit,
wherein the processor circuit is arranged to obtain an input image and metadata,
wherein the input image has pixels,
wherein the pixels have input luminances,
wherein the input luminances fall within an input luminance dynamic range,
wherein the input luminance dynamic range has an input maximum luminance,
wherein the metadata comprises a reference luminance mapping function;
wherein the processor circuit is arranged to determine an adapted luminance mapping function,
wherein the adapted luminance mapping function is based on the reference luminance mapping function,
wherein the reference luminance mapping function specifies a relationship between luminances of the input image and reference luminances,
wherein the reference luminances have a reference maximum luminance,
wherein the processor circuit is arranged to specify for each point on a diagonal, a metric along a line segment oriented in a direction,
wherein the diagonal is set in coordinate system of the input luminances normalized to a maximum of 1 and output luminances normalized to a maximum of 1,
wherein the line segment has a first endpoint on the diagonal and has a second endpoint at an intersection of the line segment and the reference luminance mapping function,
wherein the processor circuit is arranged to locate a position on each metric corresponding to an output maximum luminance;
wherein the processor circuit is arranged to calculate a histogram of intermediate luminances by applying the adapted luminance mapping function to the input luminances,
wherein the processor circuit is arranged to calculate an average brightness measure on the basis of the histogram,
wherein the processor circuit is arranged to calculate a first strength from the average brightness measure,
wherein the processor circuit is arranged to calculate a weighted sum of pixel counts in at least two brightest bins of the histogram,
wherein the processor circuit is arranged to calculate a second strength value from the weighted sum,
wherein the processor circuit is arranged to determine a booster strength based on a difference,
wherein the difference is the first strength minus the second strength,
wherein the processor circuit is arranged to calculate an adjusted adapted luminance mapping function as a locus of positions of the booster strength on the metric,
wherein the processor circuit is arranged to apply the adjusted adapted luminance mapping function to the input luminances so as to obtain the output luminances.
7 . A computer program stored on a non-transitory medium, wherein the computer program when executed on a processor performs the method as claimed in claim 1 .
8 . The apparatus as claimed in claim 6 , wherein one of the at least two brightest bins of the histogram comprises the 1% brightest pixels in the histogram.
9 . The apparatus as claimed in claim 6 , wherein the booster strength is determined as a multiplier multiplied by a relative boost.
10 . The apparatus as claimed in claim 6 , wherein the calculating of the booster strength is based on temporal filtering so as to limit a change with a previously determined booster strength.
11 . The apparatus as claimed in claim 10 ,
wherein temporal filtering is asymmetric,
wherein the temporal filtering is rapidly changing to lesser booster strength,
wherein the temporal filtering is slower changing to higher booster strength.