• 한국산업융합학회 논문집
• /
• 제23권3호
• /
• pp.515-521
• /
• 2020

Deep learning methods have been effectively used to provide great improvement in various research fields such as machine learning, image processing and computer vision. One of the most frequently used deep learning methods in image processing is the convolutional neural networks. Compared to the traditional artificial neural networks, convolutional neural networks do not use the predefined kernels, but instead they learn data specific kernels. This property makes them to be used as feature extractors as well. In this study, we compared the quality of CNN features for traditional texture feature extraction methods. Experimental results demonstrate the superiority of the CNN features. Additionally, the recognition process and result of a pioneering CNN on MNIST database are presented.

• 문화기술의 융합
• /
• 제3권4호
• /
• pp.165-169
• /
• 2017

CNN(Convolutional Nerual Network)는 기계학습 알고리즘 중에서도 이미지의 인식과 분류에 뛰어난 성능을 보이는 알고리즘 중 하나이다. CNN의 경우 간단하지만 많은 연산량을 가지고 있어 많은 시간이 소요된다. 따라서 본 논문에서는 CNN 수행과정에서 많은 처리시간이 소모되는 convolution layer와 pooling layer, fully connected layer의 연산수행을 SIMT(Single Instruction Multiple Thread)구조의 GPGPU(General-Purpose computing on Graphics Processing Units)를 통하여 병렬로 연산처리를 수행했다. 또한 convolution layer의 출력을 저장하지 않고 pooling layer의 입력으로 바로 사용함으로 메모리 접근횟수를 줄여 성능 향상을 기대했다. 본 논문에서는 이 실험검증을 위하여 MNIST 데이터 셋을 사용하였고 이를 통하여 제안하는 CNN 구조가 기존의 구조보다 12.38% 더 좋은 성능을 보임을 확인했다.

• IEIE Transactions on Smart Processing and Computing
• /
• 제6권1호
• /
• pp.53-59
• /
• 2017

Optical character recognition (OCR) automatically recognizes text in an image. OCR is still a challenging problem in computer vision. A successful solution to OCR has important device applications, such as text-to-speech conversion and automatic document classification. In this work, we analyze character recognition performance using the current state-of-the-art deep-learning structures. One is the AlexNet structure, another is the LeNet structure, and the other one is the SPNet structure. For this, we have built our own dataset that contains digits and upper- and lower-case characters. We experiment in the presence of salt-and-pepper noise or Gaussian noise, and report the performance comparison in terms of recognition error. Experimental results indicate by five-fold cross-validation that the SPNet structure (our approach) outperforms AlexNet and LeNet in recognition error.

• International Journal of Computer Science & Network Security
• /
• 제24권6호
• /
• pp.8-16
• /
• 2024

Communication with the deaf has always been crucial. Deaf and hard-of-hearing persons can now express their thoughts and opinions to teachers through sign language, which has become a universal language and a very effective tool. This helps to improve their education. This facilitates and simplifies the referral procedure between them and the teachers. There are various bodily movements used in sign language, including those of arms, legs, and face. Pure expressiveness, proximity, and shared interests are examples of nonverbal physical communication that is distinct from gestures that convey a particular message. The meanings of gestures vary depending on your social or cultural background and are quite unique. Sign language prediction recognition is a highly popular and Research is ongoing in this area, and the SVM has shown value. Research in a number of fields where SVMs struggle has encouraged the development of numerous applications, such as SVM for enormous data sets, SVM for multi-classification, and SVM for unbalanced data sets.Without a precise diagnosis of the signs, right control measures cannot be applied when they are needed. One of the methods that is frequently utilized for the identification and categorization of sign languages is image processing. African Buffalo Optimization using Support Vector Machine (ABO+SVM) classification technology is used in this work to help identify and categorize peoples' sign languages. Segmentation by K-means clustering is used to first identify the sign region, after which color and texture features are extracted. The accuracy, sensitivity, Precision, specificity, and F1-score of the proposed system African Buffalo Optimization using Support Vector Machine (ABOSVM) are validated against the existing classifiers SVM, CNN, and PSO+ANN.

• Journal of information and communication convergence engineering
• /
• 제2권2호
• /
• pp.132-137
• /
• 2004

Most of the current computer vision theories are based on hypotheses that are difficult to apply to the real world, and they simply imitate a coarse form of the human visual system. As a result, they have not been showing satisfying results. In the human visual system, there is a mechanism that processes information due to memory degradation with time and limited storage space. Starting from research on the human visual system, this study analyzes a mechanism that processes input information when information is transferred from the retina to ganglion cells. In this study, a model for the characteristics of ganglion cells in the retina is proposed after considering the structure of the retina and the efficiency of storage space. The MNIST database of handwritten letters is used as data for this research, and ART2 and SOM as recognizers. The results of this study show that the proposed recognition model is not much different from the general recognition model in terms of recognition rate, but the efficiency of storage space can be improved by constructing a mechanism that processes input information.

• 한국멀티미디어학회논문지
• /
• 제24권11호
• /
• pp.1552-1559
• /
• 2021

Convolution neural networks (CNNs) show notable performance in image processing and are used as representative core models. CNNs extract and learn features from large amounts of train dataset. In general, it has a structure in which a convolution layer and a fully connected layer are stacked. The core of CNN is the convolution layer. The size of the kernel used for feature extraction and the number that affect the depth of the feature map determine the amount of weight parameters of the CNN that can be learned. These parameters are the main causes of increasing the computational complexity and memory usage of the entire neural network. The most computationally expensive components in CNNs are fully connected and spatial convolution computations. In this paper, we propose a Fourier Convolution Neural Network that performs the operation of the convolution layer in the Fourier domain. We work on modifying and improving the amount of computation by applying the fast fourier transform method. Using the MNIST dataset, the performance was similar to that of the general CNN in terms of accuracy. In terms of operation speed, 7.2% faster operation speed was achieved. An average of 19% faster speed was achieved in experiments using 1024x1024 images and various sizes of kernels.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제10권9호
• /
• pp.375-384
• /
• 2021

딥러닝 기법은 영상 처리 분야에서 높은 성능을 입증 받아 다양한 분야에서 적용되고 있다. 이러한 딥러닝 모델의 검증에 가장 널리 사용되는 방법으로는 홀드아웃 검증 방법, k-겹 교차 검증 방법, 부트스트랩 방법 등이 있다. 이러한 기존의 기법들은 데이터 셋을 분할하는 과정에서 클래스 간의 비율에 대한 균형을 고려하지만, 같은 클래스 내에서도 존재하는 다양한 특징들의 비율은 고려하지 않고 있다. 이러한 특징들을 고려하지 않을 경우, 일부 특징에 편향된 검증 결과를 얻게 될 수 있다. 따라서 본 논문에서는 기존 검증 방법들을 개선하여 영상 분류를 위한 데이터 특징 커버리지 기반의 딥러닝 모델 검증 기법을 제안한다. 제안하는 기법은 딥러닝 모델의 학습과 검증을 위한 훈련 데이터 셋과 평가 데이터 셋이 전체 데이터 셋의 특징을 얼마나 반영하고 있는지 수치로 측정할 수 있는 데이터 특징 커버리지를 제안한다. 이러한 방식은 전체 데이터 셋의 특징을 모두 포함하도록 커버리지를 보장하여 데이터 셋을 분할할 수 있고, 모델의 평가 결과를 생성한 특징 군집 단위로 분석할 수 있다. 검증결과, 훈련 데이터 셋의 데이터 특징 커버리지가 낮아질 경우, 모델이 특정 특징에 편향되게 학습하여 모델의 성능이 낮아지며, Fashion-MNIST의 경우 정확도가 8.9%까지 차이나는 것을 확인하였다.