• Journal of Intelligence and Information Systems
• /
• v.25 no.4
• /
• pp.53-65
• /
• 2019

Object tracking is one of important steps to achieve video-based surveillance systems. Object tracking is considered as an essential task similar to object detection and recognition. In order to perform object tracking, various machine learning methods (e.g., least-squares, perceptron and support vector machine) can be applied for different designs of tracking systems. In general, generative methods (e.g., principal component analysis) were utilized due to its simplicity and effectiveness. However, the generative methods were only focused on modeling the target object. Due to this limitation, discriminative methods (e.g., binary classification) were adopted to distinguish the target object and the background. Among the machine learning methods for binary classification, total error rate minimization can be used as one of successful machine learning methods for binary classification. The total error rate minimization can achieve a global minimum due to a quadratic approximation to a step function while other methods (e.g., support vector machine) seek local minima using nonlinear functions (e.g., hinge loss function). Due to this quadratic approximation, the total error rate minimization could obtain appropriate properties in solving optimization problems for binary classification. However, this total error rate minimization was based on a batch mode setting. The batch mode setting can be limited to several applications under offline learning. Due to limited computing resources, offline learning could not handle large scale data sets. Compared to offline learning, online learning can update its solution without storing all training samples in learning process. Due to increment of large scale data sets, online learning becomes one of essential properties for various applications. Since object tracking needs to handle data samples in real time, online learning based total error rate minimization methods are necessary to efficiently address object tracking problems. Due to the need of the online learning, an online learning based total error rate minimization method was developed. However, an approximately reweighted technique was developed. Although the approximation technique is utilized, this online version of the total error rate minimization could achieve good performances in biometric applications. However, this method is assumed that the total error rate minimization can be asymptotically achieved when only the number of training samples is infinite. Although there is the assumption to achieve the total error rate minimization, the approximation issue can continuously accumulate learning errors according to increment of training samples. Due to this reason, the approximated online learning solution can then lead a wrong solution. The wrong solution can make significant errors when it is applied to surveillance systems. In this paper, we propose an exactly reweighted technique to recursively update the solution of the total error rate minimization in online learning manner. Compared to the approximately reweighted online total error rate minimization, an exactly reweighted online total error rate minimization is achieved. The proposed exact online learning method based on the total error rate minimization is then applied to object tracking problems. In our object tracking system, particle filtering is adopted. In particle filtering, our observation model is consisted of both generative and discriminative methods to leverage the advantages between generative and discriminative properties. In our experiments, our proposed object tracking system achieves promising performances on 8 public video sequences over competing object tracking systems. The paired t-test is also reported to evaluate its quality of the results. Our proposed online learning method can be extended under the deep learning architecture which can cover the shallow and deep networks. Moreover, online learning methods, that need the exact reweighting process, can use our proposed reweighting technique. In addition to object tracking, the proposed online learning method can be easily applied to object detection and recognition. Therefore, our proposed methods can contribute to online learning community and object tracking, detection and recognition communities.

• Journal of radiological science and technology
• /
• v.36 no.2
• /
• pp.123-129
• /
• 2013

This study investigated the rate of setup variance by the rotating unbalance of gantry in image-guided radiation therapy. The equipments used linear accelerator(Elekta Synergy TM, UK) and a three-dimensional volume imaging mode(3D Volume View) in cone beam computed tomography(CBCT) system. 2D images obtained by rotating $360^{\circ}$and $180^{\circ}$ were reconstructed to 3D image. Catpan503 phantom and homogeneous phantom were used to measure the setup errors. Ball-bearing phantom was used to check the rotation axis of the CBCT. The volume image from CBCT using Catphan503 phantom and homogeneous phantom were analyzed and compared to images from conventional CT in the six dimensional view(X, Y, Z, Roll, Pitch, and Yaw). The variance ratio of setup error were difference in X 0.6 mm, Y 0.5 mm Z 0.5 mm when the gantry rotated $360^{\circ}$ in orthogonal coordinate. whereas rotated $180^{\circ}$, the error measured 0.9 mm, 0.2 mm, 0.3 mm in X, Y, Z respectively. In the rotating coordinates, the more increased the rotating unbalance, the more raised average ratio of setup errors. The resolution of CBCT images showed 2 level of difference in the table recommended. CBCT had a good agreement compared to each recommended values which is the mechanical safety, geometry accuracy and image quality. The rotating unbalance of gentry vary hardly in orthogonal coordinate. However, in rotating coordinate of gantry exceeded the ${\pm}1^{\circ}$ of recommended value. Therefore, when we do sophisticated radiation therapy six dimensional correction is needed.

• Asia pacific journal of information systems
• /
• v.20 no.1
• /
• pp.81-99
• /
• 2010

Lately, in consumers' markets the number of new items is rapidly increasing at an overwhelming rate while consumers have limited access to information about those new products in making a sensible, well-informed purchase. Therefore, item providers and customers need a system which recommends right items to right customers. Also, whenever new items are released, for instance, the recommender system specializing in new items can help item providers locate and identify potential customers. Currently, new items are being added to an existing system without being specially noted to consumers, making it difficult for consumers to identify and evaluate new products introduced in the markets. Most of previous approaches for recommender systems have to rely on the usage history of customers. For new items, this content-based (CB) approach is simply not available for the system to recommend those new items to potential consumers. Although collaborative filtering (CF) approach is not directly applicable to solve the new item problem, it would be a good idea to use the basic principle of CF which identifies similar customers, i,e. neighbors, and recommend items to those customers who have liked the similar items in the past. This research aims to suggest a hybrid recommendation procedure based on the preference boundary of target customer. We suggest the hybrid recommendation procedure using the preference boundary in the feature space for recommending new items only. The basic principle is that if a new item belongs within the preference boundary of a target customer, then it is evaluated to be preferred by the customer. Customers' preferences and characteristics of items including new items are represented in a feature space, and the scope or boundary of the target customer's preference is extended to those of neighbors'. The new item recommendation procedure consists of three steps. The first step is analyzing the profile of items, which are represented as k-dimensional feature values. The second step is to determine the representative point of the target customer's preference boundary, the centroid, based on a personal information set. To determine the centroid of preference boundary of a target customer, three algorithms are developed in this research: one is using the centroid of a target customer only (TC), the other is using centroid of a (dummy) big target customer that is composed of a target customer and his/her neighbors (BC), and another is using centroids of a target customer and his/her neighbors (NC). The third step is to determine the range of the preference boundary, the radius. The suggested algorithm Is using the average distance (AD) between the centroid and all purchased items. We test whether the CF-based approach to determine the centroid of the preference boundary improves the recommendation quality or not. For this purpose, we develop two hybrid algorithms, BC and NC, which use neighbors when deciding centroid of the preference boundary. To test the validity of hybrid algorithms, BC and NC, we developed CB-algorithm, TC, which uses target customers only. We measured effectiveness scores of suggested algorithms and compared them through a series of experiments with a set of real mobile image transaction data. We spilt the period between 1st June 2004 and 31st July and the period between 1st August and 31st August 2004 as a training set and a test set, respectively. The training set Is used to make the preference boundary, and the test set is used to evaluate the performance of the suggested hybrid recommendation procedure. The main aim of this research Is to compare the hybrid recommendation algorithm with the CB algorithm. To evaluate the performance of each algorithm, we compare the purchased new item list in test period with the recommended item list which is recommended by suggested algorithms. So we employ the evaluation metric to hit the ratio for evaluating our algorithms. The hit ratio is defined as the ratio of the hit set size to the recommended set size. The hit set size means the number of success of recommendations in our experiment, and the test set size means the number of purchased items during the test period. Experimental test result shows the hit ratio of BC and NC is bigger than that of TC. This means using neighbors Is more effective to recommend new items. That is hybrid algorithm using CF is more effective when recommending to consumers new items than the algorithm using only CB. The reason of the smaller hit ratio of BC than that of NC is that BC is defined as a dummy or virtual customer who purchased all items of target customers' and neighbors'. That is centroid of BC often shifts from that of TC, so it tends to reflect skewed characters of target customer. So the recommendation algorithm using NC shows the best hit ratio, because NC has sufficient information about target customers and their neighbors without damaging the information about the target customers.

• The Korean Journal of Nuclear Medicine
• /
• v.36 no.6
• /
• pp.344-356
• /
• 2002

Purpose: RBC blood pool SPECT has been used to diagnose focal liver lesion such as hemangioma owing to its high specificity. However, low spatial resolution is a major limitation of this modality. Recently, ordered subset expectation maximization (OSEM) has been introduced to obtain tomographic images for clinical application. We compared this new modified iterative reconstruction method, OSEM with conventional filtered back projection (FBP) in imaging of liver hemangioma. Materials and Methods: Sixty four projection data were acquired using dual head gamma camera in 28 lesions of 24 patients with cavernous hemangioma of liver and these raw data were transferred to LINUX based personal computer. After the replacement of header file as interfile, OSEM was performed under various conditions of subsets (1,2,4,8,16, and 32) and iteration numbers (1,2,4,8, and 16) to obtain the best setting for liver imaging. The best condition for imaging in our investigation was considered to be 4 iterations and 16 subsets. After then, all the images were processed by both FBP and OSEM. Three experts reviewed these images without any information. Results: According to blind review of 28 lesions, OSEM images revealed at least same or better image quality than those of FBP in nearly all cases. Although there showed no significant difference in detection of large lesions more than 3 cm, 5 lesions with 1.5 to 3 cm in diameter were detected by OSEM only. However, both techniques failed to depict 4 cases of small lesions less than 1.5 cm. Conclusion: OSEM revealed better contrast and define in depiction of liver hemangioma as well as higher sensitivity in detection of small lesions. Furthermore this reconstruction method dose not require high performance computer system or long reconstruction time, therefore OSEM is supposed to be good method that can be applied to RBC blood pool SPECT for the diagnosis of liver hemangioma.

• The Korean Journal of Nuclear Medicine
• /
• v.34 no.4
• /
• pp.344-352
• /
• 2000

Purpose: Re-188-Hydroxyethylidene diphosphonate (HEDP) is a new cost-effective agent for systemic radioisotope therapy of metastatic bone pain. We investigated the influence of carrier for labeling and biodistribution of Re-188-HEDP using HEDP kit with or without carrier ($KReO_4$). Materials and Methods: The kits (HEDP 15 mg, gentisic acid 4 mg and $SnCl_2.2H_2O_2$ 4.5 mg) with or without carrier ($KReO_4$ 0.1 mg) were labeled with Re-188 solution, made available from an in-house generator by boiling for 15 min. We compared the labeling efficiency and stability of carrier-added and carrier-free preparations of Re-188-HEDP Biodistribution and imaging studies of each preparation were performed in ICR mice ($1.85{\sim}3.7MBq/0.1ml$) and SD rats ($74.1{\sim}85.2MBq/0.5ml$). Results: The carrier-added preparation showed high labeling efficiency (95% at pH 5) and high stability in serum (88%, 3 hr). However, the carrier-free preparation showed low labeling efficiency (59% at pH 5) and low stability (43%, 3 hr). The carrier-added preparation showed high uptake in bone and low uptake in stomach and kidneys. However, the carrier-free preparation showed lower uptake in bone and higher uptake in both stomach and kidneys, which is supposed to be due to released perrhenate. The carrier-added preparation also showed better images with higher skeletal accumulation, lower uptake in other organs and lower soft tissue uptake than the carrier-free preparation. Conclusion: The results of these studies clearly demonstrate that addition of carrier perrhenate is required for high labeling efficiency, stability, bone uptake and good image quality of Re-188-HEDP.

• Journal of The Korean Society of Grassland and Forage Science
• /
• v.36 no.2
• /
• pp.89-97
• /
• 2016

This study was carried out to determine the effect of mixed pasture using domestic varieties orchardgrass 'Kodione' and tall fescue 'Purumi' on forage yields and botanical composition from 2013 to 2015 in middle region of Korea. The following mixed pastures were used in this study: treatment A (orchardgrass oriented mixture with imported varieties), treatment B (orchardgrass oriented mixture), treatment C (orchardgrass and tall fescue oriented mixture), treatment D (tall fescue oriented mixture), treatment E (tall fescue oriented mixture without orchardgrass), treatment F (only orchardgrass), and treatment G (only tall fescue). In botanical compositions at last cutting time in 2015, treatment A had orchardgrass at 7.1%, tall fescue at 47.9%, kentucky bluegrass at 15.2% and white clover at 25%. Meanwhile, treatment B had orchardgrass at 25.9%, tall fescue at 47.0%, kentucky bluegrass at 4.9% and white clover at 18.3%. Consequentially, in botanical composition from 2014 to 2015, treatment A was changed more compare to other treatments B, C, D and E. The average of dry matter (DM) yield for 2 years of treatment E (18,369 kg/ha) was the highest among the seven treatments, but there was no significant difference among other treatments except treatment A (p>0.05). The crude protein (CP) and in vitro dry matter digestibility (IVDMD) of treatment A were 14.5% and 74.8%, and treatment E were 14.1% and 73.0%, respectively. The content of neutral detergent fiber (NDF) and acid detergent fiber (ADF) were lower in treatment E and A than other treatments, and treatment E showed high content of total digestible nutrient (TDN). Therefore, tall fescue oriented mixture using domestic tall fescue variety Purumi had good forage productivity, quality and botanical composition in middle region of Korea for the establishment of grassland.

• Journal of The Korean Society of Grassland and Forage Science
• /
• v.36 no.2
• /
• pp.142-149
• /
• 2016

'Shinseong', a winter forage triticale cultivar (X Triticosecale Wittmack), was developed at the Department of Rice and Winter Cereal Crop, NICS, RDA in 2015. The cultivar 'Shinseong' was selected from the cross RONDO/2$^*$ERIZO_11//KISSA_4/3/ASNO/ARDI_3//ERIZO_7 by CIMMYT (Mexico) in 1998. Subsequent generations were handled in pedigree selection programs at Mexico from 1999 to 2004, and a line 'CTSS98Y00019S-0MXI-B-3-3-5' was selected for earliness and good agronomic characteristics. After preliminary and advance yield testing in Korea for 3 years, the line was designated 'Iksan47'. The line was subsequently evaluated for earliness and forage yield in seven locations, Jeju, Iksan, Cheongwon, Yesan, Gangjin, Daegu, and Jinju from 2013 to 2015 and was finally named 'Shinseong'. Cultivar 'Shinseong' has the characteristics of light green leaves, yellow culm and spike, and a medium grain of brown color. The heading date of cultivar 'Shinseong' was April 24 which was 3 days earlier than that of check cultivar 'Shinyoung'. The tolerance or resistance to lodging, wet injury, powdery mildew, and leaf rust of 'Shinseong' were similar to those of the check cultivar. The average forage dry matter yield of cultivar 'Shinseong' at milk-ripe stages was $15MT\;ha^{-1}$, which was 3% lower than that ($15.5MT\;ha^{-1}$) of the check cultivar 'Shinyoung'. The silage quality of 'Shinseong' (6.7%) was higher than that of the check cultivar 'Shinyoung' (5.9%) in crude protein content, while was similar to the check cultivar 'Shinyoung' in acid detergent fiber (34.6%), neutral detergent fiber (58.6%), and total digestible nutrients (61.6%). It showed grain yield of $7.2MT\;ha^{-1}$ which was 25% higher than that of the check cultivar 'Shinyoung' ($5.8MT\;ha^{-1}$). This cultivar is recommended for fall sowing forage crops in areas in which average daily minimum mean temperatures in January are higher than $-10^{\circ}C$.

• Korean journal of food and cookery science
• /
• v.28 no.4
• /
• pp.413-421
• /
• 2012

This study was conducted with 418 adults 20 years or older, all of whom lived in Daegu and Gyeongbuk. According to a survey, the number of females was greater than that of males (40.7%) by 59.3%, and that for ages 30 years was the highest. The preference for Korean traditional drinks was relatively high at 51.8%, and the frequency of drinking Korean traditional drinks was 39.0%. The adults answered that they had these beverages on special days such as holidays, ritual days, and birthdays. Among the reasons for drinking a traditional beverage 'good taste' scored the highest with 27.0% of respondents, followed by 'Korean traditional food' with 24.4%. The recognition of Korean traditional drinks was high in the order of Sikhe, Soojunggwa, Cha, and Hwachae. The preference for Sikhe was the highest. The group who agreed that it was important to develop a Korean functional traditional drink was 11.5% higher than that of the negative group, as 13.4% 'agreed a lot' and 41.1% 'agreed'. Consumer awareness toward traditional drink functionality was generally positive, with 3.5 points or higher on average, and awareness of the nutritional supplementation, diabetes control, the recuperative effects of the drinks were also high. Among Korean traditional drinks Sikhe was the highest with regard to intention to purchase. As a result, the popularization of traditional Korean drinks was based on three factors: quality oriented image, popularity oriented image, and product attribute-oriented image. These factors significantly influenced the preference for and purchase of Korean functional traditional drinks.

• Korean journal of food and cookery science
• /
• v.19 no.2
• /
• pp.165-173
• /
• 2003

This study was carried out to develop a better fruit-dressing than previous mayonnaise in terms of calories, salt content and color, using kiwis, oranges, strawberries, and apples. This study was aimed to identify the perception, physico-chemical and sensory characteristics of the 4-kinds of fruits dressing. The results were as follows: The perception of the fruit dressings received the highest score (3.90) for 'It's convenience foods'. The misture in the dressings ranged from 49.3 to 59.4%, and was highest in apple dressing, followed by the kiwi, strawberry, and orange dressings, in that order. The crude ash for the dressings ranged from 0.8 to 1.4%, and was highest in the orange dressing. The L value was highest in the apple dressing, the a(-) value in the kiwi dressing, the a(＋) value in the strawberry dressing and the b value in the orange dressing. The Ca/ P ratio was about 1.17 : 1. The results of mineral contents analyses were also examined. The calcium, magnesium and potassium levels were highest in the orange dressing. The mineral contents of fruit dressings were 6.7 to 20.4mg/100g for calcium, 6.5 to 20.5mg/100g for magnesium and 67.2 to 182.6mg/100g for potassium. The sodium content of the fruit dressings were lower than in the previous mayonnaise, especially for the kiwi-dressing, which was 30% that in the mayonnaise. As a result of the sensory evaluations of the 4-kinds of fruits dressing, the kiwi dressing received the highest scores (3.50) of all the fruit-dressings in relation to the taste and palatability. Fruit are good for producing dressings that are low in calories and salt, with kiwi dressing proving to be the best of those fruits tested.

• Korean Journal of Food Science and Technology
• /
• v.10 no.3
• /
• pp.320-330
• /
• 1978

Riced dehydrated potato granules with a good reconstituton quality was made from Irish Cobber (IC) and Shimabara which are representative varieties in Korea. A part of IC potato granules was packed in small vinyl-bags while the rest, including some BHA (75ppm)-treated granules, was packed in $301{\times}407$ plain tin cans. These granules were stored in a room at room temperature, and their physico-chemical properties and lipid composition were studied. The color of the granules was measured with a Hunter-type Tristimules colorimeter. L, $a_L,$ and $b_L-values$ were respectively 83.8, -1.1 and 18.3. Variety and package-type did not affect the color. In general, it was darker than that of commercial wheat flour, but whiter than that of commercial defatted soy flour. At $100^{\circ}C$, the granules, irrespective of variety and package-type, absorbed water rapidly, and reached the maximum moisture content of $90{\sim}92%$ in 3 min., whereas they absorbed water more slowly at $8{\sim}14^{\circ}C$, and reached the maximum content of $72{\sim}74%$ in 5 minutes. Peroxide, TBA, carbonyl, acid, and iodine values of the granules after 3 months storage were respetively $150{\sim}460\;meq/kg,\;20{\sim}26,\;154{\sim}380$ micromole, $24{\sim}59,\;and\;70{\sim}78$. Except iodine values, all the chemical values were affected by variety and package-type. Ether and $CHCl_3-extracts$ of the granules from IC and Shimabara were subjected to GLC and TLC analyses. In case of IC, the major fatty acids were palmitic, stearic, and oleic (30.0, 18.8, 40.6), while in case of Shimabara they were palmitic, oleic, and linoleic acids (26.7, 39.6, 23.4%). The major lipid classes were, in both cases, triglycerides and phospholipids. Their contents were raspectively 19.1 and 43.1 (IC), and 30.1 and 37.4% (Shimabara).