Detection of pulmonary nodules with ground glass opacity (GGO) is a difficult task in radiology. Follow up is often required in medical fields. But diagnosis based on CT images are dependent on ability and experience of radiologists. In addition to that, enormous number of images increase their burden. So, to improve the detection accuracy and to reduce the burden of doctors, a CAD (Computer Aided Diagnosis) system is expected. So, in this paper, we propose an automatic algorithm for GGO detection on CT images. At first, vessel areas are removed from original CT images by using 3D Line Filter and then candidate regions are detected by threshold processing. After that, we calculate statistical features of segmented candidate regions and use artificial neural network (ANN) to distinguish final candidate regions. We applied the proposed method to 31 CT image sets in the Lung Image Database Consortium (LIDC) which is supplied by National Center Institute (NCI). In this paper, we show the experimental results and give discussions.
Most of workpiece shapes in NC milling simulations are in Z-map representations that require a very large amount of data to precisely hold a high resolution model. An irreversible compression algorithm for Z-map models using a two-dimensional Haar wavelet transform is proposed to resolve this tight memory situation for an ordinary PC. A shape model is first transformed by using Haar wavelet to build a wavelet synopsis tree while the maximum errors caused by virtually truncating high-frequency components are simultaneously calculated. The total amount of the shape data can be reduced by truncating particular sections of the wavelet components that satisfy the error threshold given by the user. Our algorithm guarantees that any error due to its irreversible compression processes is smaller than the specified level measured against the original model. A series of experiments were conducted using an Apple iMac with a 3.2 GHz CPU and 8 GB of memory. The experiments were performed with 16 sample shape models on 512×512 to 8192×8192 grids to evaluate the compression efficiency of the proposed method. Experimental results confirmed that our compression algorithm requires approximately 20–30 ms for 512×512 models and 7 s for 8192×8192 models under a maximum error level of 10× 10−6 m (a typical criteria for NC milling simulations). The compressed binaries outputted by the proposed method are generally 25–35% smaller than the baseline results by gzip, one of common reversible compression libraries, while these two methods require almost the same level of computational costs.
Highlights Discarding diagonal components in WT significantly reduces data amount. The proposed method outperforms a reversible method by 25–35% in size reduction. Most of computational time is consumed by the reversible compression step. The proposed method compresses 5122 models in 20 ms, 81922 models in 7 s.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.