Skip to main content
Fig. 1 | Journal of Analytical Science and Technology

Fig. 1

From: Deep learning image segmentation for the reliable porosity measurement of high-capacity Ni-based oxide cathode secondary particles

Fig. 1

Microstructure of high-capacity Ni-based oxide cathodes (LiNi0.8Co0.1Mn0.1O2, NCM). a SEM surface image of a secondary particle of NCM. b SEM–FIB serial section structure of a secondary particle of NCM. c High-magnification backscatter electron (BSE) image obtained from the FIB cross-sectioning workflow. Note that statistical noise background and vertical stripe contrast artifacts in all the acquired BSE images were removed using Gaussian and FFT filters before constructing the image stack. d Experimental BSE image stack with a cropped size of 256 × 256 pixels and e human-based binarized pore image stack prepared as ground truth datasets after applying hysteresis thresholding. f Concept of the gradient magnitude of pixel intensity to determine the hysteresis thresholding condition

Back to article page