hi pallini, Source package is having 2 source files 1) picojpeg.c and jpg2tga.c For your reerence I am attaching the link again here ">http://code.google.com/p/picojpeg/" thanks srini

Hi pallini, I have used visual C++ 2008 express edition to execute the code and works fine, Since we are using Kiel as our IDE I did some minor changes. Yes only one header file(picojpeg.h) and they have used malloc in jpg2tga.c file

hi CPallini, Actually for using malloc we need to allocate heap space in startup.s file, here its allowing only 1000k of memory and if I enter 2000k it will throw error as "there is no space in execution region", so when it try to access the memory at 35k it will show error, below is the code where it is getting stuck. pImage = (uint8 *)malloc(row_pitch * decoded_height);// here the size of allocation is >32kb. if (!pImage) { fclose(g_pInFile); return NULL; } row_blocks_per_mcu = image_info.m_MCUWidth >> 3; col_blocks_per_mcu = image_info.m_MCUHeight >> 3; for ( ; ; ) { int y, x; uint8 *pDst_row; status = pjpeg_decode_mcu(); if (status) { if (status != PJPG_NO_MORE_BLOCKS) { printf("pjpeg_decode_mcu() failed with status %u\n", status); free(pImage); fclose(g_pInFile); return NULL; } break; } if (mcu_y >= image_info.m_MCUSPerCol) { free(pImage); fclose(g_pInFile); return NULL; } if (reduce) { // In reduce mode, only the first pixel of each 8x8 block is valid. pDst_row = pImage + mcu_y * col_blocks_per_mcu * row_pitch + mcu_x * row_blocks_per_mcu * image_info.m_comps; if (image_info.m_scanType == PJPG_GRAYSCALE) { *pDst_row = image_info.m_pMCUBufR[0]; } else { uint y, x; for (y = 0; y < col_blocks_per_mcu; y++) { uint src_ofs = (y * 128); for (x = 0; x < row_blocks_per_mcu; x++) { pDst_row[0] = image_info.m_pMCUBufR[src_ofs]; // Here the error occurs. pDst_row[1] = image_info.m_pMCUBufG[src_ofs]; pDst_row[2] = image_info.m_pMCUBufB[src_ofs]; pDst_row += 3; src_ofs += 64; } pDst_row += row_pitch - 3 * row_blocks_per_mcu; } } } mcu_x++; if (mcu_x == image_info.m_MCUSPerRow) { mcu_x = 0; mcu_y++; } } fclose(g_pInFile); *x = decoded_width; *y = decoded_height; *comps = image_info.m_comps; return pImage; thanks srini

Pallini, They said the code memory consumption 2.3kb, not the code uses. actually while resizing an image of 1MB file, it reduces the size by taking the first pixel of each 8x8 block, so for that they will allocate the size of >32 kb of memory. thanks srini

Hi all, I Sreenivasa, new to embedded and I am writing a application that needs to re size an image(Jpeg format) and these are my requirements: * C * Keil as ide. * Image format is JPEG to JPEG. So for I tried with the open source code in the below link http://code.google.com/p/picojpeg/ It works fine but it needs more than 32K of memory allocation, but my MCU is having only 32k of RAM. Any help is appreciated. thanks Srini