Previous research presented the structured illumination confocal scanning microscope (SICSM) so as to improve the lateral resolution of the confocal microscope. However, the image acquisition speed ...of the SICSM is very slow and also an alignment error due to the mechanical rotation of a grating and a slit can easily occur. As a theoretical study, in this paper we propose a new SI method, the cross SI method, which improves lateral resolution and image acquisition speed. Performances of the conventional SI and the proposed SI methods are compared by analysis of the modulation transfer function. The proposed SI method shows similar lateral resolution and can shorten the image acquisition time compared to the conventional SI method. The cross structured illumination confocal microscope (CSICM) is combined with the cross SI pattern optics and the line scanning confocal microscope. We have introduced a 2-D diffractive grating, four linear polarizers and four cylindrical lenses in order to create the cross SI pattern. The effects of the cross SI pattern, intensity and visibility, on the system performance are analyzed. The CSICM has double the lateral resolution of the conventional microscope, an optical sectioning ability and a fast image acquisition speed.
Foreign substances in food cause disgust to consumers and some cases directly harm their health. Therefore, the detection of foreign substances in the food production process is very important, and ...active research has been conducted to date. In the case of the RGB image-based foreign substance detection system currently used in industrial sites, the accuracy is low when detecting foreign substances that are difficult to distinguish in the visible spectrum. Besides, it is difficult to detect foreign substances having similar color and texture to seaweed. In this paper, we propose a method for detecting foreign substances in seaweed using VNIR (Visible and Near-Infrared) hyperspectral images. The VNIR hyperspectral image is characterized by dividing the wavelength from the visible spectrum to the near-infrared spectrum very finely, and the camera used in the experiment has a total of 224 spectral characteristics per pixel. Spectral analysis of the acquired hyperspectral image enables more sophisticated foreign substance detection than the conventional method, and the advantage is that accurate location information and shape information can be obtained through pixel-based detection. Through the experiment, it has proven that it is possible to detect foreign substances that are difficult to distinguish with the naked eye, such as foreign substances having a similar color as seaweed or very small (about 1mm) foreign substances.
Foreign objects in food can cause disgust in consumers as well as have a direct impact on health. With the recent development of image recognition technology using deep learning, many studies are ...being conducted to detect foreign objects in food through deep learning. Deep learning can learn features well in roughly uniform distributions of class labels. However, the classes of foreign objects are diverse and difficult to collect industrial site. As a result, there is a problem with the distribution of long-tailed data with a large number of normal classes and a few abnormal classes. Moreover, even though deep learning, adjacent objects are difficult to classify because their boundaries are ambiguous. In this study, we focus on finding foreign objects overlapped to the green onion flakes that are the base material used in many countries. To detect foreign objects (e.g. insect, hair, etc.) overlapped to green onion flakes, we develop artificial minority over-sampling method. Through this method, training data is generated for foreign objects overlapped to green onion flakes. Our network classified images of foreign objects overlapped to green onion flakes 94.29% success ratio among a total of 105 objects. The results show that when trained with the proposed re-sampling, the network is able to achieve significant performance gains on foreign objects overlapped to green onion flakes.
Transaction Level Modeling (TLM) approach is used to meet the simulation speed as well as cycle accuracy for large scale SoC performance analysis. We implemented a transaction-level model of a ...proprietary bus called AHB+ which supports an extended AMBA2.0 protocol. The AHB+ transaction-level model shows 353 times faster than pin-accurate RTL model while maintaining 97% of accuracy on average. We also present the development procedure of TLM of a bus architecture.
It is difficult to evaluate R&D performance especially for companies’ R&D in various industries because each industry has different characteristics and specific industrial environments. R&D ...performance evaluation can be measured by its quantitative output and firms’ performance. In this paper, R&D performance evaluation for each industry is analyzed by the indices of quantitative output and outcome, and the differences among industries can be checked for the indices. That is, one index can stand out in an industry, but another one can be remarkable in other industries. Whether the differences for each index among industries exist can be checked. We surveyed four industries such as machinery, electronics, parts & materials and IT by scoring index method with weight factors. The survey was conducted for small & medium-sized enterprises (SMEs) in Korea focused on R&D performance evaluation indices classified by output and outcome result which are gained from several hundred firms. More sensitive indices are specified for each industry, and the indices can be used with higher priority for evaluating firm’s R&D performance. By further research of the differences for each industry, the weight for indices can be specified in each industry, and the evaluation among heterogeneous industries can be performed by normalizing indices.