02604nam a2200193 a 450000100080000000500110000800800410001910000190006024501090007926002850018830000140047352018220048765000120230965000090232170000170233070000210234770000220236870000200239021784672025-09-02       2014    bl uuuu     u00u1 u    #d1 aSCHEUER, P. M.  aImage analysis tools for characterization of whole wheat bread with fat replacer.h[electronic resource]  aIn: CONGRESSO BRASILEIRO DE CIÊNCIA E TECNOLOGIA DE ALIMENTOS, 24., 2014, Aracajú; CONGRESSO DO INSTITUTO NACIONAL DE CIÊNCIA E TECNOLOGIA DE FRUTOS TROPICAIS, 4., 2014, Aracajú. Inovação e sustentabilidade em ciência e tecnologia de alimentos: anais. Aracajú: SBCTAc2014  c1 CD-ROM.  aA healthy diet can help prevent chronic diseases, especially if it considers using whole grains and fewer calories from fat components. Several dietary guidelines suggest eating plenty of whole grain cereal foods without added fat. This work aimed to compare two different image thresholding techniques to quantify differences in bread crumb structure made with 60% whole wheat flour and 3% fat replacer, performed by baking test Official Method of the AACC, with adaptations. The crumb cell analysis was performed by scanning 1-cm-thick longitudinal and cross section samples of three slices bread. The images were saved as bmp, 300 dpi, RGB color. A single 40 mm square field of view was evaluated for each image. The free GNU Image Manipulation Program was used to make the measurements. Pre-processing (change to greylevel), segmentation (manual thresholding tool with greyscale polarization into binary images, using Otsu and Default algorithms) and crumb grain measurements (extraction of parameters) of the images were carried out using the ImageJ software. The bread crumb properties evaluate d were: number of cells (objects); mean cell area (mm2); cell density (nº cell/mm2); circularity. ) between the crumb bread image analysis parameters, number of cells (628.7 and 759.0 objects), mean cell area (0.298 and 0.434 mm2), cell density (2280.5 and 1740.9 nº cell/mm2) and circularity (0.833 and 0.819)) using Otsu and Default algorithms, respectively. The numerous crumb cells were characterized by a small cell area &lt 0.1 mm2 with both tools (Otsu, 97% and Default, 93%). The difference between the two thresholding methods was less than 4% units and this value seemed to decrease as the cell size increased. The thresholding methods applied (Otsu and Default) rendered similar results. Funder: FAPESC.  aGordura  aPão1 aMATTIONI, B.1 aFRANCISCO, A. de1 aMIRANDA, M. Z. de1 aBAYER, V. M. L.