• 지능정보연구
• /
• 제15권2호
• /
• pp.49-59
• /
• 2009

내용기반 이미지 검색은 색상, 질감 등의 이미지 자체의 자질들을 이용하여 검색하므로 텍스트 기반 이미지 검색의 객관성 부족과 모든 이미지에 사람이 주석을 달아야 하는 단점을 보완할 수 있는 이미지 검색 방법이다. 이러한 내용 기반 이미지 검색에서 사용되는 방식 중 SIM(Self-organizing Image browsing Map) 방식은 SOM 알고리즘을 이용하여 이미지들을 브라우징 가능한 그룹으로 맵핑하고 그 결과를 바탕으로 이미지를 검색하게 된다. 하지만 비슷한 이미지라 할지라도 이미지의 밝기, 피사체의 움직임 등에 의하여 색상 정보가 다르게 나타나게 되면 SOM 알고리즘의 학습 과정에서 유사한 이미지들을 그룹화한 노드를 BMU로 선택하지 못하고 떨어져 있는 다른 노드를 선택하게 된다. 이 경우 학습이 진행되면서 유사한 이미지들이 군집하는 과정을 거치지만 학습이 완료될 때까지 다른 유사 이미지들을 그룹화한 노드에 맵핑이 되지 못하는 경우가 발생한다. 그 결과, 검색 결과에 나타나지 못하여 적합 이미지 검색률이 낮아 질 수 있다. 따라서 본 논문에서는 HSV 색상모델을 이용하여 양자화하고 이미지의 색상 특징 벡터를 추출한 뒤 SOM 알고리즘을 이용하여 이미지들을 브라우징 가능한 그룹으로 맵핑한다. 이때 SIM 방식의 문제점인 유사 이미지가 따로 맵핑되어 적합 이미지 검색률이 낮아지는 것을 줄이기 위하여 SOM을 두 개의 층으로 구성한다. 첫 번째 층에서 이미지의 색상 자질을 이용하여 학습을 완료한 후, 학습이 완료된 첫 번째 층 맵의 각 노드들의 연결 가중치를 이용하여 두 번째 층에서 다시 한번 학습을 수행한다. 두 개의 층으로 학습이 완료된 두 번째 층의 SOM에 질의 이미지의 특징 벡터를 입력하여 BMU를 선택하고 BMU와 연결된 첫 번째 층의 노드를 최종 선택하여 이미지를 검색한다. 실험결과, 제안된 이미지 검색 방법이 기존의 이미지 검색 방법 보다 적합 이미지의 검색 성공률이 높은 것을 확인 할 수 있었다.

• Korean Journal of Radiology
• /
• 제24권4호
• /
• pp.294-304
• /
• 2023

Objective: We aimed to investigate whether image standardization using deep learning-based computed tomography (CT) image conversion would improve the performance of deep learning-based automated hepatic segmentation across various reconstruction methods. Materials and Methods: We collected contrast-enhanced dual-energy CT of the abdomen that was obtained using various reconstruction methods, including filtered back projection, iterative reconstruction, optimum contrast, and monoenergetic images with 40, 60, and 80 keV. A deep learning based image conversion algorithm was developed to standardize the CT images using 142 CT examinations (128 for training and 14 for tuning). A separate set of 43 CT examinations from 42 patients (mean age, 10.1 years) was used as the test data. A commercial software program (MEDIP PRO v2.0.0.0, MEDICALIP Co. Ltd.) based on 2D U-NET was used to create liver segmentation masks with liver volume. The original 80 keV images were used as the ground truth. We used the paired t-test to compare the segmentation performance in the Dice similarity coefficient (DSC) and difference ratio of the liver volume relative to the ground truth volume before and after image standardization. The concordance correlation coefficient (CCC) was used to assess the agreement between the segmented liver volume and ground-truth volume. Results: The original CT images showed variable and poor segmentation performances. The standardized images achieved significantly higher DSCs for liver segmentation than the original images (DSC [original, 5.40%-91.27%] vs. [standardized, 93.16%-96.74%], all P < 0.001). The difference ratio of liver volume also decreased significantly after image conversion (original, 9.84%-91.37% vs. standardized, 1.99%-4.41%). In all protocols, CCCs improved after image conversion (original, -0.006-0.964 vs. standardized, 0.990-0.998). Conclusion: Deep learning-based CT image standardization can improve the performance of automated hepatic segmentation using CT images reconstructed using various methods. Deep learning-based CT image conversion may have the potential to improve the generalizability of the segmentation network.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2003년도 ISIS 2003
• /
• pp.480-483
• /
• 2003

We propose a new framework that aims at multi-purpose image recognition, a difficult task for the conventional rule-based systems. This framework is farmed based on the idea of computer-based learning algorithm. In this research, we introduce the new functions of an additional learning and a knowledge reconstruction on the Fuzzy Inference Neural Network (FINN) (1) to enable the system to accommodate new objects and enhance the accuracy as necessary. We examine the capability of the proposed framework using two examples. The first one is the capital letter recognition task from UCI machine learning repository to estimate the effectiveness of the framework itself, Even though the whole training data was not given in advance, the proposed framework operated with a small loss of accuracy by introducing functions of the additional learning and the knowledge reconstruction. The other is the scenery image recognition. We confirmed that the proposed framework could recognize images with high accuracy and accommodate new object recursively.

• Journal of information and communication convergence engineering
• /
• 제16권3호
• /
• pp.173-178
• /
• 2018

Accurate classification of cloud images is a challenging task. Almost all the existing methods rely on hand-crafted feature extraction. Their limitation is low discriminative power. In the recent years, deep learning with convolution neural networks (CNNs), which can auto extract features, has achieved promising results in many computer vision and image understanding fields. However, deep learning approaches usually need large datasets. This paper proposes a deep learning approach for classification of cloud image patches on small datasets. First, we design a suitable deep learning model for small datasets using a CNN, and then we apply data augmentation and dropout regularization techniques to increase the generalization of the model. The experiments for the proposed approach were performed on SWIMCAT small dataset with k-fold cross-validation. The experimental results demonstrated perfect classification accuracy for most classes on every fold, and confirmed both the high accuracy and the robustness of the proposed model.

• 대한임베디드공학회논문지
• /
• 제13권6호
• /
• pp.313-319
• /
• 2018

We propose a post-processing algorithm to improve the quality of the RGB image generated by deep learning based colorization from the gray-scale image of an infrared camera. Wavelet fusion is used to generate a new luminance component of the RGB image luminance component from the deep learning model and the luminance component of the infrared camera. PSNR is increased for all experimental images by applying the proposed algorithm to RGB images generated by two deep learning models of SegNet and DCGAN. For the SegNet model, the average PSNR is improved by 1.3906dB at level 1 of the Haar wavelet method. For the DCGAN model, PSNR is improved 0.0759dB on the average at level 5 of the Daubechies wavelet method. It is also confirmed that the edge components are emphasized by the post-processing and the visibility is improved.

• 센서학회지
• /
• 제30권2호
• /
• pp.76-81
• /
• 2021

In this study, we propose a machine vision system with a high object recognition rate. By utilizing a multiple-exposure image sensing technique, the proposed deep learning-based machine vision system can cover a wide light intensity range without further learning processes on the various light intensity range. If the proposed machine vision system fails to recognize object features, the system operates in a multiple-exposure sensing mode and detects the target object that is blocked in the near dark or bright region. Furthermore, short- and long-exposure images from the multiple-exposure sensing mode are synthesized to obtain accurate object feature information. That results in the generation of a wide dynamic range of image information. Even with the object recognition resources for the deep learning process with a light intensity range of only 23 dB, the prototype machine vision system with the multiple-exposure imaging method demonstrated an object recognition performance with a light intensity range of up to 96 dB.

• International Journal of Advanced Culture Technology
• /
• 제9권1호
• /
• pp.190-195
• /
• 2021

Currently, one of the fields leading the 4th industrial revolution is the image recognition field of artificial intelligence, which is showing good results in many fields. In this paper, using is a YOLO V2 model, which is one of the image recognition models, we intend to classify and select into three types according to the characteristics of fruits. To this end, it was designed to proceed the number of iterations of learning 9000 counts based on 640 mandarin image data of 3 classes. For model evaluation, normal, rotten, and unripe mandarin oranges were used based on images. We as a result of the experiment, the accuracy of the learning model was different depending on the number of learning. Normal mandarin oranges showed the highest at 60.5% in 9000 repetition learning, and unripe mandarin oranges also showed the highest at 61.8% in 9000 repetition learning. Lastly, rotten tangerines showed the highest accuracy at 86.0% in 7000 iterations. It will be very helpful if the results of this study are used for fruit farms in rural areas where labor is scarce.

• 센서학회지
• /
• 제33권3호
• /
• pp.119-124
• /
• 2024

Effective fire sensing is important to protect lives and property from the disaster. In this paper, we present an intelligent visual sensing method for detecting fires based on machine learning techniques. The proposed method involves a two-step process. In the first step, fire and non-fire images are used to train a convolutional neural network (CNN), and in the next step, feature vectors consisting of 256 values obtained from the CNN are used for the learning of a support vector machine (SVM). Linear and nonlinear SVMs with different parameters are intensively tested. We found that the proposed hybrid method using an SVM with a linear kernel effectively increased the recall rate of fire image detection without compromising detection accuracy when an imbalanced dataset was used for learning. This is a major contribution of this study because recall is important, particularly in the sensing of disaster situations such as fires. In our experiments, the proposed system exhibited an accuracy of 96.9% and a recall rate of 92.9% for test image data.

• Advances in nano research
• /
• 제12권2호
• /
• pp.185-195
• /
• 2022

Deep learning is another field of artificial intelligence (AI) utilized for computer aided diagnosis (CAD) and image processing in scientific research. Considering numerous mechanical repetitive tasks, reading image slices need time and improper with geographical limits, so the counting of image information is hard due to its strong subjectivity that raise the error ratio in misdiagnosis. Regarding the highest mortality rate of Lung cancer, there is a need for biopsy for determining its class for additional treatment. Deep learning has recently given strong tools in diagnose of lung cancer and making therapeutic regimen. However, identifying the pathological lung cancer's class by CT images in beginning phase because of the absence of powerful AI models and public training data set is difficult. Convolutional Neural Network (CNN) was proposed with its essential function in recognizing the pathological CT images. 472 patients subjected to staging FDG-PET/CT were selected in 2 months prior to surgery or biopsy. CNN was developed and showed the accuracy of 87%, 69%, and 69% in training, validation, and test sets, respectively, for T1-T2 and T3-T4 lung cancer classification. Subsequently, CNN (or deep learning) could improve the CT images' data set, indicating that the application of classifiers is adequate to accomplish better exactness in distinguishing pathological CT images that performs better than few deep learning models, such as ResNet-34, Alex Net, and Dense Net with or without Soft max weights.

• 로봇학회논문지
• /
• 제16권1호
• /
• pp.56-63
• /
• 2021

As the use of robots in service area increases, research has been conducted to replace human tasks in daily life with robots. Among them, this study focuses on the tidy-up task on a desk using a robot arm. The order in which tidy-up motions are carried out has a great impact on the success rate of the task. Therefore, in this study, a neural network-based method for determining the priority of the tidy-up motions from the input image is proposed. Reinforcement learning, which shows good performance in the sequential decision-making process, is used to train such a task planner. The training process is conducted in a virtual tidy-up environment that is configured the same as the actual tidy-up environment. To transfer the learning results in the virtual environment to the actual environment, the input image is preprocessed into a segmented image. In addition, the use of a neural network that excludes unnecessary tidy-up motions from the priority during the tidy-up operation increases the success rate of the task planner. Experiments were conducted in the real world to verify the proposed task planning method.