• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2023.05a
• /
• pp.127-128
• /
• 2023

The construction industry is increasingly adopting vision AI technologies to improve efficiency and safety management. However, the complex and dynamic nature of construction sites can pose challenges to the accuracy of vision AI models trained on datasets that do not consider the background. This study investigates the effect of background on object recognition for vision AI in construction sites by constructing a learning dataset and a test dataset with varying backgrounds. Frame scaffolding was chosen as the object of recognition due to its wide use, potential safety hazards, and difficulty in recognition. The experimental results showed that considering the background during model training significantly improved the accuracy of object recognition.

• Journal of Digital Contents Society
• /
• v.16 no.4
• /
• pp.605-613
• /
• 2015

This paper addresses the problem to make sparse the projection matrix in pattern recognition method. Recently, the size of computer program is often restricted in embedded systems. It is very often that developed programs include some constant data. For example, many pattern recognition programs use the projection matrix for dimension reduction. To improve the recognition performance, very high dimensional feature vectors are often extracted. In this case, the projection matrix can be very big. Recently, RSR(roated sparse regression) method[1] was proposed. This method has been proved one of the best algorithm that obtains the sparse matrix. We propose three methods to improve the RSR; outlier removal, sampling and elastic net RSR(E-RSR) in which the penalty term in RSR optimization function is replaced by that of the elastic net regression. The experimental results show that the proposed methods are very effective and improve the sparsity rate dramatically without sacrificing the recognition rate compared to the original RSR method.

• KIPS Transactions on Software and Data Engineering
• /
• v.5 no.5
• /
• pp.251-260
• /
• 2016

Face recognition is a technology to extract feature from a facial image, learn the features through various algorithms, and recognize a person by comparing the learned data with feature of a new facial image. Especially, in order to improve the rate of face recognition, face recognition requires various processing methods. In the training stage of face recognition, feature should be extracted from a facial image. As for the existing method of extracting facial feature, linear discriminant analysis (LDA) is being mainly used. The LDA method is to express a facial image with dots on the high-dimensional space, and extract facial feature to distinguish a person by analyzing the class information and the distribution of dots. As the position of a dot is determined by pixel values of a facial image on the high-dimensional space, if unnecessary areas or frequently changing areas are included on a facial image, incorrect facial feature could be extracted by LDA. Especially, if a camera image is used for face recognition, the size of a face could vary with the distance between the face and the camera, deteriorating the rate of face recognition. Thus, in order to solve this problem, this paper detected a facial area by using a camera, removed unnecessary areas using the facial feature area calculated via a Gabor filter, and normalized the size of the facial area. Facial feature were extracted through LDA using the normalized facial image and were learned through the artificial neural network for face recognition. As a result, it was possible to improve the rate of face recognition by approx. 13% compared to the existing face recognition method including unnecessary areas.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.5
• /
• pp.388-402
• /
• 2003

In this paper, we carried out the word, 4 continuous digits. continuous, and task-independent word recognition experiments to verify the effectiveness of the re-defined phoneme-likely units (PLUs) for the phonetic decision tree based HM-Net (Hidden Markov Network) context-dependent (CD) acoustic modeling in Korean appropriately. In case of the 48 PLUs, the phonemes /ㅂ/, /ㄷ/, /ㄱ/ are separated by initial sound, medial vowel, final consonant, and the consonants /ㄹ/, /ㅈ/, /ㅎ/ are also separated by initial sound, final consonant according to the position of syllable, word, and sentence, respectively. In this paper. therefore, we re-define the 39 PLUs by unifying the one phoneme in the separated initial sound, medial vowel, and final consonant of the 48 PLUs to construct the CD acoustic models effectively. Through the experimental results using the re-defined 39 PLUs, in word recognition experiments with the context-independent (CI) acoustic models, the 48 PLUs has an average of 7.06%, higher recognition accuracy than the 39 PLUs used. But in the speaker-independent word recognition experiments with the CD acoustic models, the 39 PLUs has an average of 0.61% better recognition accuracy than the 48 PLUs used. In the 4 continuous digits recognition experiments with the liaison phenomena. the 39 PLUs has also an average of 6.55% higher recognition accuracy. And then, in continuous speech recognition experiments, the 39 PLUs has an average of 15.08% better recognition accuracy than the 48 PLUs used too. Finally, though the 48, 39 PLUs have the lower recognition accuracy, the 39 PLUs has an average of 1.17% higher recognition characteristic than the 48 PLUs used in the task-independent word recognition experiments according to the unknown contextual factor. Through the above experiments, we verified the effectiveness of the re-defined 39 PLUs compared to the 48PLUs to construct the CD acoustic models in this paper.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.18 no.3
• /
• pp.373-380
• /
• 2008

So far there have been many researches on machine-based recognition of handwritten digit. But we have not yet attained the level of performance that can be satisfactory to men. The dissatisfaction with the performance of machine comes from not only the low accuracy of recognition but also the dissimilarity of the recognition results between man and machine. To reduce the difference of machine from man we first made an experiment with the human recognition of handwritten digits and then inquiry into the way of the human recognition that makes the results of men different from that of machine. We found out the attributes that play an important role in the human recognition process through the analysis of the experimental results like uni- and bi-directional confused pairs of digits, several ones unmixed up with another and the redundancy of mis-recognition, and proposed the approach direction to be able to improve the accuracy of the machine-based recognition, and furthermore the similarity in the recognition results of men and machine on the basis of the found facts above.

• Convergence Security Journal
• /
• v.23 no.5
• /
• pp.3-8
• /
• 2023

The Fourth Industrial Revolution is bringing about great changes in many industrial fields, and among them, active research is being conducted on convergence technology using artificial intelligence. Among them, the demand is increasing day by day in the field of object recognition using artificial intelligence and digital transformation using recognition results. In this paper, we proposed an optimal learning model construction method to accurately recognize letters, symbols, and lines in images and save the recognition results as files in a standardized format so that they can be used in simulations. In order to recognize letters, symbols, and lines in images, the characteristics of each recognition target were analyzed and the optimal recognition technique was selected. Next, a method to build an optimal learning model was proposed to improve the recognition rate for each recognition target. The recognition results were confirmed by setting different order and weights for character, symbol, and line recognition, and a plan for recognition post-processing was also prepared. The final recognition results were saved in a standardized format that can be used for various processing such as simulation. The excellent performance of building the optimal learning model proposed in this paper was confirmed through experiments.

• The Journal of the Acoustical Society of Korea
• /
• v.10 no.3
• /
• pp.31-36
• /
• 1991

We propose an efficient pre-classification algorithm extracting candidate words to reduce the recognition time in a large-vocabulary recognition system and also propose the use of spectral and temporal smoothing of the observation probability to improve its classification performance. The proposed algorithm computes the coarse likelihood score for each word in a lexicon using the observation probabilities of speech spectra and duration information of recognition units. With the proposed approach we could reduce the computational amount by 74% with slight degradation of recognition accuracy in 1160-word recognition system based on the phoneme-level HMM. Also, we observed that the proposed coarse likelihood score computation algorithm is a good estimator of the likelihood score computed by the Viterbi algorithm.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.3
• /
• pp.175-181
• /
• 2003

Because a word of Korean digits is a syllable and deeply coarticulatied in connected digits, some recognition models based on demisyllables have been proposed by researchers. However, they could not show an excellent recognition results yet. This paper proposes a recognition model based on extended and context-dependent demisyllables, such as a tri-demisyllable like a tri-phone, for the Korean 4-connected digits recognition. For experiments, we use a toolkit of HTK 3.0 for building this model of continuous HMMs using training Korean connected digits from SiTEC database and for recognizing unknown ones. The results show that the recognition rate is 92% and this model has an ability to improve the recognition performance of Korean connected digits.

• Fire Science and Engineering
• /
• v.24 no.5
• /
• pp.10-25
• /
• 2010

As disaster potential power of modern society grows larger, to improve and reinforce efficiently a national system which prepares and responds disasters, analyzed the survey for government officials of the department disaster management. Following is the contents of this research. First, cooperative relationship to disaster management organizations. Second, necessity of law establishment related crisis and disaster department. Third, by recognition regarding disaster management situational variable, overall recognition regarding disaster management situation, overall recognition regarding crisis type, recognition regarding occurrence possibility along disaster scale. Fourth, by recognition regarding structural variable of disaster management, the National Emergency Management Agency regarding disaster management, related organization, recognition difference of local government. It is a research about confusion regarding step of prevention - preparation - correspondence - restoration.

• International journal of advanced smart convergence
• /
• v.9 no.1
• /
• pp.113-120
• /
• 2020

Recently, the license plate OCR system has been commercialized in a variety of fields and preferred utilizing low-cost embedded systems using only cameras. This system has a high recognition rate of about 98% or more for the environments such as parking lots where non-vehicle is restricted; however, the environments where non-vehicle objects are not restricted, the recognition rate is about 50% to 70%. This low performance is due to the changes in the environment by non-vehicle objects in real-time situations that occur anomaly data which is similar to the license plates. In this paper, we implement the appropriate anomaly detection based on semi-supervised learning for the license plate OCR system in the real-time environment where the appearance of non-vehicle objects is not restricted. In the experiment, we compare systems which anomaly detection is not implemented in the preceding research with the proposed system in this paper. As a result, the systems which anomaly detection is not implemented had a recognition rate of 77%; however, the systems with the semi-supervised learning based on anomaly detection had 88% of recognition rate. Using the techniques of anomaly detection based on the semi-supervised learning was effective in detecting anomaly data and it was helpful to improve the recognition rate of real-time situations.