Method and apparatus for selecting a coding mode

Abstract

A method and apparatus for selecting a coding mode are described. For example, the method receives at least one block of a signal to be encoded. The method determines a frame vertical pixel difference in the at least one block and determines a field vertical pixel difference in the at least one block. The method then compares the frame vertical pixel difference with the field vertical pixel difference to determine a first coding mode for the at least one block.

What is claimed is: 
     
       1. A method for selecting a frame/field coding mode per frame picture, the method comprising:
 receiving a frame picture of a signal to be encoded; 
 determining a sum of an absolute frame vertical pixel difference over the frame picture according to:
   Σ abs|Xi,j−Xi,j+ 1|,
 
 
 wherein i and j are a pixel horizontal and vertical indices, X i,j  represents a first line, and Xi,j+1 represents a next line of a same frame picture, such that the sum of the absolute frame vertical pixel difference is a sum of differences between lines in the same frame picture; 
 determining a sum of an absolute field vertical pixel differences over the same frame picture according to:
   Σ abs|Xi, 2 j−Xi, 2( j+ 1)|+Σ abs|Xi, 2 j+ 1− Xi, 2( j+ 1)+1|,
 
   wherein 
   Σ abs|Xi, 2− Xi, 2( j+ 1)|
 
 
 represents a difference between even field lines in the same frame picture and
   Σ abs|Xi, 2 j+ 1− Xi, 2( j+ 1)+1|
 
 
 represents a difference between odd field lines in the same frame picture; comparing the sum of the absolute frame vertical pixel differences with the sum of the absolute field vertical pixel difference; 
 selecting a field coding mode for encoding the frame picture if the sum of the frame vertical pixel difference is greater than the sum of the field vertical pixel differences, else 
 selecting a frame coding mode for encoding the frame picture. 
 
     
     
       2. The method of  claim 1 , wherein the at least one picture comprises a partition comprising at least one of: a 16×16 block, a 16×8 block, an 8×16 block, an 8×8 block, an 8×4 block, a 4×8 block, or a 4×4 block. 
     
     
       3. A method for selecting a frame/field coding mode per intra picture, the method comprising:
 performing intra prediction for blocks within the intra picture in both frame and field coding modes; 
 determine a minimum cost for each of the blocks in the frame coding mode; 
 determining a minimum cost for each of the blocks in the field coding mode; 
 determining a minimum frame coding cost for the intra picture by summing the minimum frame costs of the blocks within the intra picture; 
 determining a minimum field coding cost for the intra picture by summing the minimum field costs of the blocks within the intra picture; 
 selecting a frame coding mode for encoding the intra picture if the minimum frame cost of the picture is less than the minimum field cost of the intra picture, else selecting the field coding mode for the intra picture. 
 
     
     
       4. A method for selecting a coding mode per inter picture, the method comprising:
 performing motion estimation (ME) in both frame and field coding modes for blocks within the inter picture; 
 determining a minimum ME cost for each of the macroblocks in the inter picture in both the frame and field coding mode;
 determining a minimum ME cost for each of the blocks in the field coding mode; and 
 determining a minimum frame coding cost for the inter picture by summing all of the minimum frame ME costs of the blocks over the inter picture; 
 determining a minimum field coding cost for inter picture by summing all of the minimum field ME costs of the blocks over the inter picture; 
 
 selecting a frame coding mode for encoding the inter picture if the minimum frame cost is less than the minimum field cost, else selecting the field coding mode for the inter picture. 
 
     
     
       5. A method for selecting a coding mode per macroblock pair or picture, the method comprising:
 identifying either a field coding mode or a frame coding mode for the macroblock pair or picture based on a vertical pixel difference comparison by:
 determining a sum of an absolute frame vertical pixel difference (Δ FRM ) over the at least one macroblock pair or picture, 
 determining a sum of an absolute field vertical pixel difference (Δ FLD ) over the macroblock pair or picture, and 
 comparing the sum of the absolute frame vertical pixel difference with the sum of the absolute field vertical pixel difference; 
 identifying a field coding mode for encoding the at least one macroblock pair or picture pair if the frame vertical pixel difference is greater than the field vertical pixel difference, else identifying a frame coding mode for encoding the at least one macroblock pair or picture pair; 
 
 identifying either a field coding mode or a frame coding mode for the macroblock pair or picture based on a ME cost by:
 for a macroblock pair, determining a minimum cost for each of the macroblocks in the macroblock pair in both the frame coding mode (J min     —     FRM ) and field coding modes (J min     —     FLD ); 
 for a picture, determining a minimum frame coding cost for all of the macroblocks in each of the macroblock pairs over the entire I picture by summing all of the minimum frame ME costs over the I picture and determining a minimum field coding cost for all of the macroblocks in each of the macroblock pairs over the entire I picture by summing all of the minimum field ME costs over the I picture; and 
 identifying a frame coding mode for encoding the macroblock pair if the minimum frame cost of the macroblock pair is less than the minimum field cost of the macroblock pair, else identifying the field coding mode for the macroblock pair; 
 
 selecting the identified coding mode for encoding the macroblock pair or picture if the same coding mode is identified based on both the vertical pixel difference comparison and the ME cost, else if 
 
       
         
           
             
               
                 
                    
                   
                     
                       J 
                       min_FRM 
                     
                     - 
                     
                       J 
                       min_FLD 
                     
                   
                    
                 
                 
                   
                     J 
                     min_FRM 
                   
                   + 
                   
                     J 
                     min_FLD 
                   
                 
               
               > 
               
                 α 
                 ⁢ 
                 
                   
                      
                     
                       
                         Δ 
                         FRM 
                       
                       - 
                       
                         Δ 
                         FLD 
                       
                     
                      
                   
                   
                     
                       Δ 
                       FRM 
                     
                     + 
                     
                       Δ 
                       FLD 
                     
                   
                 
               
             
           
         
         select the coding mode identified based on the ME cost, or if 
       
       
         
           
             
               
                 
                    
                   
                     
                       J 
                       min_FRM 
                     
                     - 
                     
                       J 
                       min_FLD 
                     
                   
                    
                 
                 
                   
                     J 
                     min_FRM 
                   
                   + 
                   
                     J 
                     min_FLD 
                   
                 
               
               < 
               
                 α 
                 ⁢ 
                 
                   
                      
                     
                       
                         Δ 
                         FRM 
                       
                       - 
                       
                         Δ 
                         FLD 
                       
                     
                      
                   
                   
                     
                       Δ 
                       FRM 
                     
                     + 
                     
                       Δ 
                       FLD 
                     
                   
                 
               
             
           
         
         select the coding mode identified based on the vertical pixel difference comparison.