Article
Computer Science, Artificial Intelligence
Yunxia Lin, Songcan Chen
Summary: Euler K-means (EulerK) maps data onto the unit hyper-sphere surface using a complex mapping, and then applies the k-means method. It is robust to noises and outliers, but the centroids captured by EulerK deviate from the unit hyper-sphere surface. We propose two Rectified Euler K-means methods (REK1 and REK2) to rectify this deviation and obtain real centroids on the mapped space for better characterization of the data structures.
PATTERN RECOGNITION
(2023)
Article
Computer Science, Information Systems
Abhash Kumar, Ashish Kumar Bhandari, Reman Kumar
Summary: This paper proposes a novel method for enhancing image contrast while retaining naturalness, using compensated histogram equalization and blending with an adaptive brightness adjustment kernel. The method is applicable to contrast degraded images and performs better than existing methods on the CSIQ dataset.
MULTIMEDIA SYSTEMS
(2021)
Article
Multidisciplinary Sciences
Canlin Li, Jinjuan Zhu, Lihua Bi, Weizheng Zhang, Yan Liu
Summary: This paper proposes a new method for enhancing low-light images by balancing brightness and preserving details. The method involves bidirectional processing of the image, including double histogram equalization and total variation model extraction. The experimental results show that the proposed method outperforms existing methods in terms of visual effect.
Article
Computer Science, Information Systems
Sanjay Agrawal, Rutuparna Panda, P. K. Mishro, Ajith Abraham
Summary: A novel joint histogram equalization (JHE) based technique is proposed in this research to improve the contrast of an image by utilizing the information among each pixel and its neighbors. The experimental analysis shows that this method outperforms the state-of-the-art histogram equalization algorithms in contrast enhancement, even for images with a narrow dynamic range.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Fei Li, Jiangbin Zheng, Yuan-fang Zhang
Summary: This paper proposes an effective generative adversarial network structure for enhancing low-light image quality. The method utilizes residual blocks and enhancing blocks to improve feature diversity and employs a loss function to recover contextual and local details. Experimental results demonstrate that the method performs well in dealing with low-light scenarios.
IET IMAGE PROCESSING
(2021)
Article
Multidisciplinary Sciences
Zheng-Jie Huang, Brijesh Patel, Wei-Hao Lu, Tz-Yu Yang, Wei-Cheng Tung, Vytautas Bucinskas, Modris Greitans, Yu-Wei Wu, Po Ting Lin
Summary: This study presents a novel approach using deep learning techniques for automatic cell detection. By optimizing image contrast and introducing a universal contrast enhancement variable, the proposed method achieves high accuracy in yeast cell detection. Comparative experiments demonstrate the superior performance of this method in cell detection, with significant improvements compared to conventional methods.
SCIENTIFIC REPORTS
(2023)
Article
Optics
Jingchun Zhou, Yanyun Wang, Weishi Zhang, Chongyi Li
Summary: This paper presents an underwater image restoration method based on feature priors, which aims to address color distortion and poor contrast issues in underwater images. By estimating background light and refining the transmission map of color corrected images, the method effectively improves the overall image quality, achieving superior performance compared to state-of-the-art methods in diverse degradation scenarios.
Article
Computer Science, Software Engineering
Roberto M. Dyke, Kai Hormann
Summary: This paper presents a novel technique that improves the naive approach of histogram equalization by linearly interpolating the cumulative distribution of a low-bit image. The proposed method is capable of producing a high entropy equalized histogram while preserving distances between similar intensities.
Article
Multidisciplinary Sciences
Yongqiang Chen, Chenglin Wen, Weifeng Liu, Wei He
Summary: In this paper, we propose an illumination enhancement network based on Retinex theory to brighten low-light images in a fast and accurate manner. The network consists of two parts: a decomposition network for separating the input image into reflectance and illumination, and an enhancement network for brightness enhancement and reflection denoising. The proposed framework utilizes cascaded iterative lighting learning and unsupervised training losses to improve illumination estimation and generalization ability. Experimental results demonstrate the effectiveness and superiority of the proposed method compared to classical and state-of-the-art methods.
SCIENTIFIC REPORTS
(2023)
Article
Construction & Building Technology
Dan Zheng, Shuaishuai Tan, Xinxin Li, Haonan Cai
Summary: Using sunlight as the heat source, defects in concrete structures were successfully identified and the detection depth was increased with the use of an enhancement algorithm. Image preprocessing methods were proposed to eliminate the effects of environment and view angle, providing a reference for infrared detection technology of concrete under weak heat sources.
ADVANCES IN CIVIL ENGINEERING
(2021)
Article
Computer Science, Information Systems
Xin Fan, Junyan Wang, Haifeng Wang, Changgao Xia
Summary: A histogram-constrained and contrast-tunable HE technique for digital image enhancement is proposed in this paper, which partitions the input image histogram into two parts and redistributes them to achieve more accurate results in terms of information entropy and MS-SSIM compared to other algorithms.
Article
Automation & Control Systems
Kun Ren, Qingyang Tao, Honggui Han
Summary: This paper proposes a novel object detection network for low-light conditions on embedded platforms, including a lightweight low-light enhancement network DS-PyLENet and an anchor-free lightweight object detector CFEDet. The testing results demonstrate that the new model outperforms the assembled model of EnlightenGAN and YOLOv5n in terms of detection accuracy and speed.
JOURNAL OF THE FRANKLIN INSTITUTE-ENGINEERING AND APPLIED MATHEMATICS
(2023)
Article
Optics
Xiuwen Bi, Pengfei Wang, Tao Wu, Fusheng Zha, Peng Xu
Summary: This paper proposes an efficient and simple method for enhancing uneven underwater illumination by exploiting the complementarity of event cameras and standard cameras, resulting in enhanced images with better scene details and colors similar to standard frames.
Article
Optics
Weitao Deng, Lei Liu, Huateng Chen, Xiaofeng Bai
Summary: This paper proposes a contrast enhancement method based on adaptive histogram correction and equalization to effectively reduce artifacts and insufficient local detail enhancement caused by traditional methods.
Article
Computer Science, Information Systems
Samer Hameed Majeed, Nor Ashidi Mat Isa
Summary: The adaptive entropy index histogram equalization (AEIHE) technique, a local subclass of HE-based contrast enhancement techniques, divides images into three sub-images using different contextual regions and clip limits to enhance and highlight local details. By combining enhanced sub-images, AEIHE ensures the production of excellent resultant images with improved contrast and highlighted local details.
Article
Computer Science, Artificial Intelligence
Ugurhan Kutbay, Firat Hardalac
Article
Computer Science, Information Systems
Firat Hardalac, Huseyin Yasar, Anil Akyel, Ugurhan Kutbay
MULTIMEDIA TOOLS AND APPLICATIONS
(2020)
Article
Engineering, Biomedical
Ugurhan Kutbay
JOURNAL OF MEDICAL AND BIOLOGICAL ENGINEERING
(2020)
Article
Computer Science, Information Systems
Ugurhan Kutbay
MULTIMEDIA TOOLS AND APPLICATIONS
(2020)
Article
Chemistry, Analytical
Firat Hardalac, Fatih Uysal, Ozan Peker, Murat Ciceklidag, Tolga Tolunay, Nil Tokgoz, Ugurhan Kutbay, Boran Demirciler, Fatih Mert
Summary: This study aims to perform fracture detection using deep-learning on wrist X-ray images to support physicians in diagnosing fractures, particularly in emergency services. The research achieved the highest detection result of 0.8639 average precision (AP50) in the developed ensemble model 'wrist fracture detection-combo (WFD-C)'. Huawei Turkey R&D Center supports this study in collaboration with Gazi University and Medskor.
Article
Engineering, Biomedical
Anil Akyel
Summary: This study proposes a novel model for stroke risk estimation. The model shows excellent accuracy and is able to accurately estimate the risk of different types of stroke.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2022)
Article
Energy & Fuels
Shukri Mahmood Younus Younus, Ugurhan Kutbay, Javad Rahebi, Firat Hardalac
Summary: This work proposes a hybrid gray wolf optimization and proportional integral controller for better speed control of Brush-less DC motors. The proposed controller is compared with PID controller, PSO-PI controller, and ANFIS under different operating settings, and it outperforms all the other controllers. Simulation results show the effectiveness of the hybrid GWO-PI-based controller.
Article
Engineering, Electrical & Electronic
Ugurhan Kutbay, Firat Hardalac, Pinar Akdemir Ozisik, Emre Unay, Cenk Yildirim
Summary: This research developed a system for measuring the closing force of clips during surgery. The system used a miniature load cell and a user-friendly design to accurately measure the force in real-time, providing neurosurgeons with important information for their operations.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2022)
Article
Computer Science, Artificial Intelligence
Firat Hardalac, Mustafa Aydin, Ugurhan Kutbay, Kubilay Ayturan, Anil Akyel, Atika Caglar, Bo Hai, Fatih Mert
Summary: This study developed a low-cost mobile support system to assist healthcare professionals in classifying neonatal jaundice using noninvasive image processing methods and estimating bilirubin levels through a simple regression curve. The algorithm achieved a classification accuracy of 92.5% in a test group of 40 subjects, showing promising results for accurate jaundice classification.
TURKISH JOURNAL OF ELECTRICAL ENGINEERING AND COMPUTER SCIENCES
(2021)
Proceedings Paper
Computer Science, Theory & Methods
Huseyin Yasar, Ugurhan Kutbay, Firat Hardalac
2018 4TH INTERNATIONAL CONFERENCE ON COMPUTER AND TECHNOLOGY APPLICATIONS (ICCTA)
(2018)
Proceedings Paper
Computer Science, Theory & Methods
Huseyin Yasar, Selami Serhatlioglu, Ugurhan Kutbay, Firat Hardalac
2018 4TH INTERNATIONAL CONFERENCE ON COMPUTER AND TECHNOLOGY APPLICATIONS (ICCTA)
(2018)
