Feature Extractionusing ORB-RANSAC for Face Recognition

Vinay, A.; Rao, Avani S.; Shekhar, Vinay S.; Kumar, Chandan; Murthy, K. N. Balasubramanya; Natarajan, S. P.

doi:10.1016/j.procs.2015.10.068

Cited by 22 publications

(4 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, not all detected keypoints may have reliable matches due to noise or image artifacts. To mitigate this, we use the Random Sample Consensus (RANSAC) algorithm, a robust matching technique (Vinay et al 2015 ). RANSAC iteratively selects a subset of correspondences from the set C , denoting correspondences between keypoints in I ref and I input as , and estimates the transformation matrix T that best aligns the keypoints and .…”

Section: Case Study: a Combined Image Processing And Deep Learning Mo...mentioning

confidence: 99%

ECG-Image-Kit: a synthetic image generation toolbox to facilitate deep learning-based electrocardiogram digitization

Shivashankara,

Deepanshi,

Mehri Shervedani

et al. 2024

Physiol. Meas.

View full text Add to dashboard Cite

Objective: Cardiovascular diseases are a major cause of mortality globally, and electrocardiograms (ECGs) are crucial for diagnosing them. Traditionally, ECGs are stored in printed formats. However, these printouts, even when scanned, are incompatible with advanced ECG diagnosis software that require time-series data. Digitizing ECG images is vital for training machine learning models in ECG diagnosis, leveraging the extensive global archives collected over decades. Deep learning models for image processing are promising in this regard, although the lack of clinical ECG archives with reference time-series data is challenging. Data augmentation techniques using realistic generative data models provide a solution.Approach: We introduce ECG-Image-Kit, an open-source toolbox for generating synthetic multi-lead ECG images with realistic artifacts from time-series data, aimed at automating the conversion of scanned ECG images to ECG data points. The tool synthesizes ECG images from real time-series data, applying distortions like text artifacts, wrinkles, and creases on a standard ECG paper background.Main results: As a case study, we used ECG-Image-Kit to create a dataset of 21,801 ECG images from the PhysioNet QT database. We developed and trained a combination of a traditional computer vision and deep neural network model on this dataset to convert synthetic images into time-series data for evaluation. We assessed digitization quality by calculating the signal-to-noise ratio (SNR) and compared clinical parameters like QRS width, RR, and QT intervals recovered from this pipeline, with the ground truth extracted from ECG time-series. The results show that this deep learning pipeline accurately digitizes paper ECGs, maintaining clinical parameters, and highlights a generative approach to digitization.Significance: The toolbox has broad applications, including model development for ECG image digitization and classification. The toolbox currently supports data augmentation for the 2024 PhysioNet Challenge, focusing on digitizing and classifying paper ECG images.

show abstract

Section: Case Study: a Combined Image Processing And Deep Learning Mo...mentioning

confidence: 99%

ECG-Image-Kit: a synthetic image generation toolbox to facilitate deep learning-based electrocardiogram digitization

Shivashankara,

Deepanshi,

Mehri Shervedani

et al. 2024

Physiol. Meas.

View full text Add to dashboard Cite

show abstract

“…Hence the Random Sample Consensus (RANSAC) is one of the model fitting analyst for image matching computations [24]. RANSAC were used on many image matching applications such as in image stitching [25], face recognitions [26] and in deep space exploration [27]. RANSAC will estimate the model parameter and filtered the outliers (gross error) of feature matching.…”

Section: Feature Points Matchingmentioning

confidence: 99%

Multispectral Image Matching Using SIFT and SURF Algorithm: A Review

2023

IJG

View full text Add to dashboard Cite

SIFT and SURF image matching were used in many industries such as survey and mapping, geology, medical and automotive. Multispectral sensors offered today become new challenge for researchers to study the performances of SIFT and SURF algorithms on multispectral image. Basically, multispectral image consists of more than three bands. As a result, the differences between those bands leads to nonlinear intensity between images. Both algorithm detectors using ‘blob detector’ that extracting the feature points as a key point for image matching later on. Hence, the less visibility of the feature on the multispectral images was one of the issues need to be solved. Many researchers investigate and propose a new strategy to extract and match the feature point using SIFT and SURF on multispectral image. The image fusions, combinations of different descriptors and revised or alteration of the algorithm themselves were among the approached taken by researchers in order to achieved good results.

show abstract

“…In order to solve the above problems, this paper proposes an ORB-based optimization method, which first sets a dynamic adaptive threshold, calculates the threshold size for feature extraction by the grey scale value of pixel points in the image, and uses this threshold to achieve image feature point extraction; the traditional ORB algorithm directly [7] uses the grey scale center of mass method to find the main direction of the feature point, but in the vicinity of the feature point, the grey scale value of the pixel to The contribution of the pixel's grey value to the feature point is different, the closer to the feature point, the greater the contribution, so the pixel's grey value is Gaussian weighted before calculating the main direction by the grey scale prime method, thus improving the accuracy and precision of the calculation.…”

Section: Introductionmentioning

confidence: 99%

Feature extraction based on improved ORB algorithm

Huang,

zuo,

Wang

2023

International Conference on Electronic Information Engineering and Data Processing (EIEDP 2023)

View full text Add to dashboard Cite

Feature Extractionusing ORB-RANSAC for Face Recognition

Cited by 22 publications

References 21 publications

ECG-Image-Kit: a synthetic image generation toolbox to facilitate deep learning-based electrocardiogram digitization

ECG-Image-Kit: a synthetic image generation toolbox to facilitate deep learning-based electrocardiogram digitization

Multispectral Image Matching Using SIFT and SURF Algorithm: A Review

Feature extraction based on improved ORB algorithm

Contact Info

Product

Resources

About