A5000747744- Advanced Image Processing Techniques
- Image Processing Techniques and Applications
- Image and Signal Denoising Methods
- Advanced Vision and Imaging
- Non-Invasive Vital Sign Monitoring
- Heart Rate Variability and Autonomic Control
- ECG Monitoring and Analysis
- Image Enhancement Techniques
- EEG and Brain-Computer Interfaces
- Advanced Image Fusion Techniques
- Image Retrieval and Classification Techniques
- Advanced Image and Video Retrieval Techniques
- Handwritten Text Recognition Techniques
Korea University
2018-2020
This paper reviews the NTIRE 2020 challenge on real world super-resolution. It focuses participating methods and final results. The addresses setting, where paired true high low-resolution images are unavailable. For training, only one set of source input is therefore provided along with a unpaired high-quality target images. In Track 1: Image Processing artifacts, aim to super-resolve synthetically generated image processing artifacts. allows for quantitative benchmarking approaches w.r.t....
This paper reviews the NTIRE 2020 challenge on perceptual extreme super-resolution with focus proposed solutions and results.The task was to super-resolve an input image a magnification factor ×16 based set of prior examples low corresponding high resolution images.The goal is obtain network design capable produce results best quality similar ground truth.The track had 280 registered participants, 19 teams submitted final results.They gauge state-of-the-art in single superresolution....
This paper proposes an unsupervised single-image Super-Resolution(SR) model using cycleGAN and domain discriminator to solve the problem of SR with unknown degradation unpaired dataset. In previous approaches, paired dataset is required for training assumed levels image degradation. real world applications, however, sets are typically not low high resolution pairs, but only images provided as inputs. To address problem, we introduce a cycle-in-cycle GAN based learning addition, combine...
Single image extreme Super Resolution (SR) is a difficult task as scale factor in the order of 10X or greater typically attempted. For instance, case 16x upscale an image, single pixel from low resolution gets expanded to 16x16 patch. Such attempts often result fuzzy quality and loss details reconstructed images. To handle these difficulties, we propose network architecture composed series connected blocks recurrent feedback fashions for enhanced SR reconstruction. By use network, refined...
We propose a novel remote heart rate (HR) estimation method using facial images based on video analytics. Most of previous methods have been demonstrated in well-controlled indoor environments. In contrast, this paper proposes practical analytic framework under actual driving conditions by extracting key HR inducing features. particular, when cars are driven, effective and stable becomes challenging as there many dynamic elements, such rapid illumination changes, vibrations, ambient lighting...
Driver assistance systems are a major focus of the automotive industry. Although technological functions that help drivers improving, monitoring driver state receives less attention. In this respect, human heart rate (HR) is one most important bio-signals, and it can be detected remotely using consumer-grade cameras. Based on this, video-based system HR signals proposed in paper. practical environment, very challenging due to changes illumination, vibrations, motion. order overcome these...
This paper reviews the NTIRE 2020 challenge on real world super-resolution. It focuses participating methods and final results. The addresses setting, where paired true high low-resolution images are unavailable. For training, only one set of source input is therefore provided along with a unpaired high-quality target images. In Track 1: Image Processing artifacts, aim to super-resolve synthetically generated image processing artifacts. allows for quantitative benchmarking approaches \wrt...
In this paper, we propose a novel framework for estimating driver's heart rate remotely under driving condition. First, region of interest is selected by landmark points derived from discriminative response map fitting. Feature signal then extracted estimation and refined dynamic standard deviation check to eliminate the noisy segments. Finally, estimated power spectral analysis using temporal filtered signal. We test our on video database captured real-world car Experimental results...
This paper reviews the NTIRE 2020 challenge on perceptual extreme super-resolution with focus proposed solutions and results. The task was to super-resolve an input image a magnification factor 16 based set of prior examples low corresponding high resolution images. goal is obtain network design capable produce results best quality similar ground truth. track had 280 registered participants, 19 teams submitted final They gauge state-of-the-art in single super-resolution.
We present a novel approach to OCR(Optical Character Recognition) of Korean character, Hangul. As phonogram, Hangul can represent 11,172 different characters with only 52 graphemes, by describing each character combination the graphemes. total number could overwhelm capacity neural network, existing OCR encoding methods pre-define smaller set that are frequently used. This design choice naturally compromises performance on long-tailed in distribution. In this work, we demonstrate grapheme is...