한국 경제자유구역의 개발 배경과 동인 -인천 송도 경제자유구역을 중심으로-

김준우 ( June Woo Kim ),안영진 ( Young Jin Ahn ) 한국도시지리학회 2006 한국도시지리학회지 Vol.9 No.3

광주의 지역발전이념: 이론적 유형화와 탐색적 조사

김준우 ( June Woo Kim ) 고려대학교 한국사회연구소 2009 한국사회 Vol.10 No.1

송도경제자유지구의 이상과 현실 -세계화와 국가의 관계를 중심으로-

김준우 ( June Woo Kim ) 한국경제지리학회 2004 한국경제지리학회지 Vol.7 No.2

The author made two major arguments in this paper based on a case study on Songdo Free Economic Zone in South Korea. First, the state power is still being maintained when looking at the development process of the project. Planning and development has been mainly on the hands of government officials. The project reflect the logic of the state rather than the market. And the government tries to manage globalization through Songdo project. Second, even though globalization did not bring about the decline of the state, it changed the state`s approach on spatial issues. Songdo is an articulate expression of pursuit of efficiency over equity with the rise of the globalization wave. Songdo also leaves the Korean state a task for consensus-building concerning liberalization.

골반환 손상 후 발생한 치명적인 폐색전증: 증례보고

김준우 ( Joon Woo Kim ),오창욱 ( Chang Wug Oh ),김형섭 ( Hyung Sub Kim ),김희준 ( Hee June Kim ) 대한외상학회 2014 대한외상학회지 Vol.27 No.4

Pelvic ring and acetabular fractures are major high-energy trauma injuries and are often associated with other injuries or fractures. In particular, a venous thromboembolism is an important common complication in trauma patients with pelvic or acetabular fractures. We report a case of a fatal pulmonary thromboembolism after a pelvic fracture in a 60- year-old female patient. [ J Trauma Inj 2014; 27: 224-8 ]

세계도시의 문화적 특성

김준우(KIM June Woo) 지역사회학회 2003 지역사회학 Vol.5 No.1

This paper deals with world city paradigm, which urban researchers contributed the most to the ongoing globalization debate. In a globalizing economy, the global control function of transnational corporations is being emphasized. World city specializes on this bringing forth the creation of world city such as New York, London, and Tokyo. It has been said that the cultural characteristics of world city is cosmopolitan and consumeristic. However, there has been continuing debate on whether there is one homogeneous world city culture or many different world city culture. This paper points out the possibility that Seoul's culture may be different from that of New York. It can be productionist and oriented to salaryman sentiment.

Transformer 네트워크를 이용한 음성신호 변환

김준우(June-Woo Kim),정호영(Ho-Young Jung) 한국음성학회 2020 말소리와 음성과학 Vol.12 No.3

음성 변환은 다양한 음성 처리 응용에 적용될 수 있으며, 음성 인식을 위한 학습 데이터 증강에도 중요한 역할을 할 수 있다. 기존의 방법은 음성 합성을 이용하여 음성 변환을 수행하는 구조를 사용하여 멜 필터뱅크가 중요한 파라미터로 활용된다. 멜 필터뱅크는 뉴럴 네트워크 학습의 편리성 및 빠른 연산 속도를 제공하지만, 자연스러운 음성파형을 생성하기 위해서는 보코더를 필요로 한다. 또한, 이 방법은 음성 인식을 위한 다양한 데이터를 얻는데 효과적이지 않다. 이 문제를 해결하기 위해 본 논문은 원형 스펙트럼을 사용하여 음성 신호 자체의 변환을 시도하였고, 어텐션 메커니즘으로 스펙트럼 성분 사이의 관계를 효율적으로 찾아내어 변환을 위한 자질을 학습할 수 있는 transformer 네트워크 기반 딥러닝 구조를 제안하였다. 영어 숫자로 구성된 TIDIGITS 데이터를 사용하여 개별 숫자변환 모델을 학습하였고, 연속 숫자 음성 변환 디코더를 통한 결과를 평가하였다. 30명의 청취 평가자를 모집하여 변환된 음성의 자연성과 유사성에 대해 평가를 진행하였고, 자연성 3.52±0.22 및 유사성 3.89±0.19 품질의 성능을 얻었다. Voice conversion can be applied to various voice processing applications. It can also play an important role in data augmentation for speech recognition. The conventional method uses the architecture of voice conversion with speech synthesis, with Mel filter bank as the main parameter. Mel filter bank is well-suited for quick computation of neural networks but cannot be converted into a high-quality waveform without the aid of a vocoder. Further, it is not effective in terms of obtaining data for speech recognition. In this paper, we focus on performing voice-to-voice conversion using only the raw spectrum. We propose a deep learning model based on the transformer network, which quickly learns the voice conversion properties using an attention mechanism between source and target spectral components. The experiments were performed on TIDIGITS data, a series of numbers spoken by an English speaker. The conversion voices were evaluated for naturalness and similarity using mean opinion score (MOS) obtained from 30 participants. Our final results yielded 3.52±0.22 for naturalness and 3.89±0.19 for similarity.

얼굴의 선상태선에서 발생한 비립종 1예

김준우 ( June Woo Kim ),김희정 ( Hee Jung Kim ),김미경 ( Mi Kyeong Kim ),이지연 ( Ji Yeoun Lee ),윤태영 ( Tae Young Yoon ) 대한피부과학회 2006 대한피부과학회지 Vol.44 No.5

A healthy 4-year-old boy presented with a linear erythematous eruption, along with some pearly papules on the right cheek. The distribution of the lesions corresponded to that of Blaschko?s lines. Histopathological examination showed necrotic keratinocytes and vacuolar changes on the basal layer in the epidermis, and a superficial and deep perivascular, perieccrine, and perifollicular lymphocytic infiltrate in the dermis. In addition, a small keratin-filled cyst was observed in the upper dermis. These findings were consistent with a diagnosis of milium developing on lichen striatus. We applied 0.03% topical tacrolimus ointment twice daily. On follow-up 12 months later, the linear erythematous eruption cleared, but the pearly papules persisted. (Korean J Dermatol 2006;44(5):639∼641)

악성 흑색종의 림프관 분포 및 밀도에 관한 면역조직화학적 연구

김준우 ( June Woo Kim ),이지연 ( Ji Yeoun Lee ),윤태영 ( Tae Young Yoon ) 대한피부과학회 2006 대한피부과학회지 Vol.44 No.10

Background: Cutaneous melanoma is a malignant neoplasm that has the tendency to metastasize through lymphatic vessels. However, the mechanism of lymphatic spread in malignant melanoma is not fully understood due to lack of lymphatic-specific markers. Objective: The purpose of this study was to determine the intra-and peritumoral lymphatic vessel density (LVD) using a novel monoclonal antibody D2-40, and whether increased expression of lymphatic vessel correlated with malignancy grading in a series of melanocytic lesions and prognosis of malignant melanoma. Methods: The intra-and peritumoral LVD were examined by immunohistochemistry using D2-40 antibody in a series of melanocytic lesions. Results: We found significantly higher intra-and peritumoral LVD in malignant melanoma as compared with either benign melanocytic nevus, dysplastic nevus, or melanoma in situ (p<0.05), and the intratumoral LVD was significantly related to Breslow depth (p<0.05) and the development of lymphatic metastasis (p<0.05). Conclusion: The higher intra-and peritumoral LVD in malignant melanomas suggests that melanoma cells might promote lymphangiogenesis. In addition, our data suggests that increased lymphatics in melanoma is an independent prognostic factor and important for the development of lymphatic metastasis. (Korean J Dermatol 2006;44(10): 1210~1216)

얼굴에 발생한 우췌상 암종

김준우 ( June Woo Kim ),김영기 ( Young Gi Kim ),김미경 ( Mi Kyeong Kim ),김지연 ( Chi Yeon Kim ),윤태영 ( Tae Young Yoon ) 대한피부과학회 2005 대한피부과학회지 Vol.43 No.4

제한된 학습 데이터를 사용하는 End-to-End 음성 인식 모델

김준우(June-Woo Kim),정호영(Ho-Young Jung) 한국음성학회 2020 말소리와 음성과학 Vol.12 No.4

음성 인식은 딥러닝 및 머신러닝 분야에서 활발히 상용화 되고 있는 분야 중 하나이다. 그러나, 현재 개발되고 있는 음성 인식 시스템은 대부분 성인 남녀를 대상으로 인식이 잘 되는 실정이다. 이것은 음성 인식 모델이 대부분 성인남녀 음성 데이터베이스를 학습하여 구축된 모델이기 때문이다. 따라서, 노인, 어린이 및 사투리를 갖는 화자의 음성을 인식하는데 문제를 일으키는 경향이 있다. 노인과 어린이의 음성을 잘 인식하기 위해서는 빅데이터를 구축하는 방법과 성인 대상 음성 인식 엔진을 노인 및 어린이 데이터로 적응하는 방법 등이 있을 수 있지만, 본 논문에서는 음향적 데이터 증강에 기반한 재귀적 인코더와 언어적 예측이 가능한 transformer 디코더로 구성된 새로운 end-to-end 모델을 제안한다. 제한된 데이터셋으로 구성된 한국어 노인 및 어린이 음성 인식을 통해 제안된 방법의 성능을 평가한다. Speech recognition is one of the areas actively commercialized using deep learning and machine learning techniques. However, the majority of speech recognition systems on the market are developed on data with limited diversity of speakers and tend to perform well on typical adult speakers only. This is because most of the speech recognition models are generally learned using a speech database obtained from adult males and females. This tends to cause problems in recognizing the speech of the elderly, children and people with dialects well. To solve these problems, it may be necessary to retain big database or to collect a data for applying a speaker adaptation. However, this paper proposes that a new end-to-end speech recognition method consists of an acoustic augmented recurrent encoder and a transformer decoder with linguistic prediction. The proposed method can bring about the reliable performance of acoustic and language models in limited data conditions. The proposed method was evaluated to recognize Korean elderly and children speech with limited amount of training data and showed the better performance compared of a conventional method.