(a) BSE image of garnet mica schist, raw version, original resolution 1024 Â 1024 pixels, 256 gray levels, original size approx. 1.3 Â 1.3 mm. Bright, large mineral phase at center is garnet. (b) Same as (a), classified version, 5 pixel classes. In terms of rising brightness: quartz, muscovite, chlorite, garnet, ilmenite. 

Contexts in source publication

Context 1

... file size and data distortion by high-loss compression. The test image is a 1024 Â 1024 pixel wide, 8 bits per pixel backscattered electron (BSE) image of a garnet mica schist that was subjected to 3-D reconstruction by serial lapping (compare Fig. 2, upper left sample). Two versions are considered: the raw BSE image with 256 gray levels ( Fig. 3a) and the classified version with five pixel classes (Fig. ...

Context 2

... The test image is a 1024 Â 1024 pixel wide, 8 bits per pixel backscattered electron (BSE) image of a garnet mica schist that was subjected to 3-D reconstruction by serial lapping (compare Fig. 2, upper left sample). Two versions are considered: the raw BSE image with 256 gray levels ( Fig. 3a) and the classified version with five pixel classes (Fig. ...

Context 3

... the continuous gray-level test image in Fig. 3a, GIF, TIFF and PNG all yield file sizes of approximately 50% of the non-compressed image size of 1MB. JPEG, using the highest image quality setting (quantization coefficient=1), yields a slightly higher compression; drastic reduction is achieved using the minimum quality setting, the resulting file size being approximately 5% of the ...

Context 4

... the highest image quality setting (quantization coefficient=1), yields a slightly higher compression; drastic reduction is achieved using the minimum quality setting, the resulting file size being approximately 5% of the non-compressed size. Clearly, file size reduction by high-loss compression sacrifices image detail. Fig. 4 shows a subarea from Fig. 3a in GIF and JPEG (medium quality) formats. In contrast to Fig. 4a, Fig. 4b shows severely blurred rims and bright artifacts around the quartz inclusions, and minor variation in the garnet area has been averaged out completely. Classifying this JPEG image with the segmentation parameters applied to Fig. 3b would result in ilmenite rims ...

Context 5

... image detail. Fig. 4 shows a subarea from Fig. 3a in GIF and JPEG (medium quality) formats. In contrast to Fig. 4a, Fig. 4b shows severely blurred rims and bright artifacts around the quartz inclusions, and minor variation in the garnet area has been averaged out completely. Classifying this JPEG image with the segmentation parameters applied to Fig. 3b would result in ilmenite rims around quartz inclusions. The no-loss LZW compression yields the best results with classified images (Fig. 3b) that include larger homogeneous areas: the GIF compressed image yields the smallest file ...

Context 6

... blurred rims and bright artifacts around the quartz inclusions, and minor variation in the garnet area has been averaged out completely. Classifying this JPEG image with the segmentation parameters applied to Fig. 3b would result in ilmenite rims around quartz inclusions. The no-loss LZW compression yields the best results with classified images (Fig. 3b) that include larger homogeneous areas: the GIF compressed image yields the smallest file ...