View Forum Topic

Forums
CS451
Black and white error.ppm solved

Author	Message
cbort Posts: 95	Posted 11:17 Jun 02, 2011 \| Ok so yesterday I had said that I would post the cause of the black and white error.ppm that some people seemed to have... ok, soooo.... pretty much this black and white is the result of a poor int to byte conversion... a byte can only handle [-128,127] and we are feeding in an int that is outside of this range causing it to loop back around... hope this helps :)
Sanaz_6101 Posts: 40	Posted 12:04 Jun 02, 2011 \| Thanks Corey for your help. Ill try to see if that would change my result :)
Sanaz_6101 Posts: 40	Posted 12:21 Jun 02, 2011 \| I have no idea why Im getting this ? see my attachment here is my mv.txt: Target Image Name IDB1/Walk_050.ppm Refrence Image Name IDB1/Walk_048.ppm Number of target macro blocks: 12x9 (Image size is: 192x144) [0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0] [0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,1][0,0] [0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0] [0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0] [0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,1] [0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0] [0,0][0,0][0,0][0,0][0,0][0,0][0,-2][-1,-3][0,0][0,0][0,0][0,0] [0,0][0,0][0,0][0,0][0,0][0,0][0,-2][0,0][0,0][0,0][0,0][0,0] [0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0][0,0] Attachments: error.ppm
cbort Posts: 95	Posted 13:19 Jun 02, 2011 \| Hmm that is pretty strange... it looks like the area where the moved object used to be is lighter than normal and darker where the object moved too... for your error are you adding it to the original image if so you shouldn't be and I think this would cause an error that looks like what you have... I haven't had anything really like that though so I cant say for certain... Were you using grayscale or YCbCR conversion on your original? is error.ppm coming from a freshly built image? what is the equation you are using to calculate the error? (it should be a simple subtraction)
Sanaz_6101 Posts: 40	Posted 14:40 Jun 02, 2011 \| No Im not adding to the original image. Also I just computer Y value, should I get CbCr as well ? here is my equation : (error[i][j] -min) * (255 ) / (max - min ) -128
cbort Posts: 95	Posted 14:46 Jun 02, 2011 \| I think you may need Cb and Cr but I am not certain... how are you calculating error[ i ] [ j ]
Sanaz_6101 Posts: 40	Posted 14:53 Jun 02, 2011 \| error [i][j] is a difference between best match and target .. my mistake , Im using this equation ( error[i][j]-min) *255 ) / (max-min) and when Im using the other one Ill get something like yours (B/W)
cbort Posts: 95	Posted 15:16 Jun 02, 2011 \| the correct equation is for Y: ((error[i][j] -min) * (255 ) / (max - min )) -128 if it lookes like than it is the problem I had mentioned to begin with I think... where is your maping being done ? it should take place for the whole image not each macro block....