35 research outputs found
불완벽한 CT 및 MR 측정 데이터로 부터의 영상 열화 보정을 위한 인공신경망에 관한 연구
No full text학위논문(박사) - 한국과학기술원 : 바이오및뇌공학과, 2019.8,[xi, 124 p. :]In this paper, we propose various reconstruction methods of medical images through a deep learning method using an artificial neural network. The various medical images discussed here include computed tomography images and magnetic resonance images. The greatest problem with the computed tomography technique is the radiation exposure to the subject. Because the amount of exposure is directly related to the safety of the subject, various studies are actively under way to reduce these levels. Among the various research methodologies, there are sparse-view imaging techniques and interior tomography techniques. The sparse-view imaging technique refers to a means of reducing the amount of exposure by intermittently measuring a reduced number of X-ray images. If a tomography image is reconstructed using this rarely acquired X-ray image, strong streaking artifacts are generated in the reconstructed image. Another method of reducing the exposure dose, the interior tomography technique, requires the localization of the x-ray imaging area to reduce exposure. If a tomography image is reconstructed using an interior X-ray image acquired by this technique, strong cupping artifacts are generated in the reconstructed image. Therefore, here an algorithm is devised to remove these streaking artifacts and cupping artifacts generated by the sparse-view CT technique and the interior tomography technique, respectively, using an artificial neural network.
With other medical images, the most serious problem with magnetic resonance imaging is that the recording time is longer. Subjects should maintain a fixed posture during the recording time in magnetic resonance imaging, which takes a few minutes to a few tens of minutes. If a subject is unable to maintain a fixed posture and moves, motion-induced noise will occur in the MRI image. In order to solve this problem, various studies are underway to develop a robust magnetic resonance imaging method or to shorten the scanning time to minimize the motion of the subject. A robust magnetic resonance imaging technique is radial trajectory imaging. The radial trajectory imaging technique shares the same mathematical theory as the sparse-view CT technique mentioned above. Therefore, the sparse-view reconstruction technique as part of the computed tomography reconstruction method is extended and applied to the reconstruction of the magnetic resonance image along the radial trajectory. A means of shortening the scanning time can be achieved by reducing the data to be acquired. Although acquisition data of magnetic resonance imaging is acquired in the frequency domain, most artificial neural network studies focus on the imaging domain. Therefore, we devised an artificial neural network that directly interpolates the frequency domain by applying an artificial neural network in the actually acquired frequency domain rather than in the existing image domain.한국과학기술원 :바이오및뇌공학과
Framing U-Net via Deep Convolutional Framelets: Application to Sparse-View CT
No full textX-ray computed tomography (CT) using sparse projection views is a recent approach to reduce the radiation dose. However, due to the insufficient projection views, an analytic reconstruction approach using the filtered back projection (FBP) produces severe streaking artifacts. Recently, deep learning approaches using large receptive field neural networks such as U-Net have demonstrated impressive performance for sparse-view CT reconstruction. However, theoretical justification is still lacking. Inspired by the recent theory of deep convolutional framelets, themain goal of this paper is, therefore, to reveal the limitation of U-Net and propose new multi-resolution deep learning schemes. In particular, we show that the alternative U-Net variants such as dual frame and tight frame U-Nets satisfy the so-called frame condition which makes them better for effective recovery of high frequency edges in sparse-view CT. Using extensive experiments with real patient data set, we demonstrate that the new network architectures provide better reconstruction performance.
One network to solve all ROIs: Deep learning CT for any ROI using differentiated backprojection
No full textDistance Sampling-based Paraphraser Leveraging ChatGPT for Text Data Manipulation
No full textThere has been growing interest in audio-language retrieval research, where
the objective is to establish the correlation between audio and text
modalities. However, most audio-text paired datasets often lack rich expression
of the text data compared to the audio samples. One of the significant
challenges facing audio-text datasets is the presence of similar or identical
captions despite different audio samples. Therefore, under many-to-one mapping
conditions, audio-text datasets lead to poor performance of retrieval tasks. In
this paper, we propose a novel approach to tackle the data imbalance problem in
audio-language retrieval task. To overcome the limitation, we introduce a
method that employs a distance sampling-based paraphraser leveraging ChatGPT,
utilizing distance function to generate a controllable distribution of
manipulated text data. For a set of sentences with the same context, the
distance is used to calculate a degree of manipulation for any two sentences,
and ChatGPT's few-shot prompting is performed using a text cluster with a
similar distance defined by the Jaccard similarity. Therefore, ChatGPT, when
applied to few-shot prompting with text clusters, can adjust the diversity of
the manipulated text based on the distance. The proposed approach is shown to
significantly enhance performance in audio-text retrieval, outperforming
conventional text augmentation techniques.Comment: Accepted at SIGIR 2024 short paper trac
Semi-supervised learning for continuous emotional intensity controllable speech synthesis with disentangled representations
No full textRecent text-to-speech models have reached the level of generating natural
speech similar to what humans say. But there still have limitations in terms of
expressiveness. The existing emotional speech synthesis models have shown
controllability using interpolated features with scaling parameters in
emotional latent space. However, the emotional latent space generated from the
existing models is difficult to control the continuous emotional intensity
because of the entanglement of features like emotions, speakers, etc. In this
paper, we propose a novel method to control the continuous intensity of
emotions using semi-supervised learning. The model learns emotions of
intermediate intensity using pseudo-labels generated from phoneme-level
sequences of speech information. An embedding space built from the proposed
model satisfies the uniform grid geometry with an emotional basis. The
experimental results showed that the proposed method was superior in
controllability and naturalness.Comment: Accepted by Interspeech 202
Fullerene-Passivated Methylammonium Lead Iodide Perovskite Absorber for High-Performance Self-Powered Photodetectors with Ultrafast Response and Broadband Detectivity
No full textWe herein report the enhanced electrical properties of self-powered perovskite-based photodetectors with high sensitivity and responsivity by applying the surface passivation strategy using C60 (fullerene) as a surface passivating agent. The perovskite (CH3NH3PbI3) thin film passivated with fullerene achieves a highly uniform and compact surface, showing reduced leakage current and higher photon-to-current conversion capability. As a result, the improved film quality of the perovskite layer allows excellent photon-detecting properties, including high values of external quantum efficiency (>95%), responsivity (>5 A W−1), and specific detectivity (>1013 Jones) at zero bias voltage, which surpasses those of the pristine perovskite-based device. Furthermore, the passivated device showed fast rise (0.18 μs) and decay times (17 μs), demonstrating high performance and ultrafast light-detecting capability of the self-powered perovskite-based photodetectors
Multi-planar 2.5D U-Net for image quality enhancement of dental cone-beam CT
No full textCone-beam computed tomography (CBCT) can provide 3D images of a targeted area with the advantage of lower dosage than multidetector computed tomography (MDCT; also simply referred to as CT). However, in CBCT, due to the cone-shaped geometry of the X-ray source and the absence of post-patient collimation, the presence of more scattering rays deteriorates the image quality compared with MDCT. CBCT is commonly used in dental clinics, and image artifacts negatively affect the radiology workflow and diagnosis. Studies have attempted to eliminate image artifacts and improve image quality; however, a vast majority of that work sacrificed structural details of the image. The current study presents a novel approach to reduce image artifacts while preserving details and sharpness in the original CBCT image for precise diagnostic purposes. We used MDCT images as reference high-quality images. Pairs of CBCT and MDCT scans were collected retrospectively at a university hospital, followed by co-registration between the CBCT and MDCT images. A contextual loss-optimized multi-planar 2.5D U-Net was proposed. Images corrected using this model were evaluated quantitatively and qualitatively by dental clinicians. The quantitative metrics showed superior quality in output images compared to the original CBCT. In the qualitative evaluation, the generated images presented significantly higher scores for artifacts, noise, resolution, and overall image quality. This proposed novel approach for noise and artifact reduction with sharpness preservation in CBCT suggests the potential of this method for diagnostic imaging.
Copyright: This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.ope
