An application of scale-invariant feature transform in iris recognition

Zhao, Weijie; Chen, Xiao; Cheng, Ji Hong; Jiang, Linhua

doi:10.1109/icis.2013.6607844

Cited by 3 publications

(4 citation statements)

References 13 publications

(27 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The matching process using Scale-Invariant Feature Transform (SIFT) consists of detection of features followed by description of the region of interest that can be used to represent uniquely during the comparison. The spatial difference between multiple resolutions of the images is explored in SIFT using Gaussian filtering and inter-pixel difference calculations (Lowe, 2004;Zhao et al, 2013).…”

Section: Resultsmentioning

confidence: 99%

Snake classification from images

James

2017

Preprint

View full text Add to dashboard Cite

Incorrect snake identification from the observable visual traits is a major reason of death resulting from snake bites. So far no automatic classification method has been proposed to distinguish snakes by deciphering the taxonomy features of snake for the two major species of snakes i.e. Elapidae and Viperidae. We present a parallel processed inter-feature product similarity fusion based automatic classification of Spectacled Cobra, Russel's Viper, King Cobra, Common Krait, Saw Scaled Viper, Hump nosed Pit Viper. We identify 31 different taxonomically relevant features from snake images for automated snake classification studies. The scalability and real-time implementation of the classifier is analyzed through GPU enabled parallel computing environment. The developed systems finds application in wild life studies, analysis of snake bites and in management of snake population.

show abstract

Section: Resultsmentioning

confidence: 99%

Snake classification from images

James

2017

Preprint

View full text Add to dashboard Cite

show abstract

“…Traditional object detection methods [33][34][35] typically employ a sliding window strategy to scan the entire image with a series of sliding windows to determine possible object locations. Hand-crafted features, such as scale-invariant feature transform 36 and histogram of oriented gradients 37 , are then extracted from the image window, followed by classification using support vector machine (SVM) or AdaBoost classifiers.However, traditional object detection algorithms based on the sliding window strategy have issues of high computational complexity, limited efficiency in detecting objects, and difficulty in handling changes in object shape and background. Additionally, designing hand-crafted features for each new object class requires considerable time.…”

Section: Traditional Approachesmentioning

confidence: 99%

Lightweight air-to-air unmanned aerial vehicle target detection model

Cheng,

Wang,

et al. 2024

Sci Rep

View full text Add to dashboard Cite

The rapid expansion of the drone industry has resulted in a substantial increase in the number of low-altitude drones, giving rise to concerns regarding collision avoidance and countermeasure strategies among these unmanned aerial vehicles. These challenges underscore the urgent need for air-to-air drone target detection. An effective target detection model must exhibit high accuracy, real-time capabilities, and a lightweight network architecture to achieve a balance between precision and speed when deployed on embedded devices. In response to these requirements, we initially curated a dataset comprising over 10,000 images of low-altitude operating drones. This dataset encompasses diverse and intricate backgrounds, significantly enhancing the model’s training capacity. Subsequently, a series of enhancements were applied to the YOLOv5 algorithm to realize lightweight object detection. A novel feature extraction network, CF2-MC, streamlined the feature extraction process, while an innovative module, MG, in the feature fusion section aimed to improve detection accuracy and reduce model complexity. Concurrently, the original CIoU loss function was replaced with the EIoU loss function to further augment the model’s accuracy. Experimental results demonstrate an enhancement in the accuracy of drone target detection, achieving mAP values of 95.4% on the UAVfly dataset and 82.2% on the Det-Fly dataset. Finally, real-world testing conducted on the Jetson TX2 revealed that the YOLOv5s-ngn model achieved an average inference speed of 14.5 milliseconds per image. The code utilized in this paper can be accessed via https://github.com/lucien22588/yolov5-ngn.git.

show abstract

“…Due to the powerful ability of feature detection in image, SIFT has been widely used in object identification, recognition, tracking, image stitching, robotic mapping and navigation [10][11][12][13]. SIFT is a feature extraction method that is invariant to image change in illumination, scale, rotation and affine.…”

Section: Scale Invariant Feature Transformmentioning

confidence: 99%

“…Besides, the existed recognition methods in II are not robust to object change in illumination, scale, rotation and affine or just robust to some of the object changes. Consequently, a new 3D object recognition method based on scale invariant feature transform (SIFT) [10][11][12][13], which is robust to object change in illumination, scale, rotation and affine, is proposed. Also, the proposed method can utilize the perspective information recorded in elemental images and don't need to search all of the depth images but can locate the exact depth image that the 3D object is included.…”

Section: Introductionmentioning

confidence: 99%

Three-dimensional object recognition via integral imaging and scale invariant feature transform

Moon

2014

SPIE Proceedings

View full text Add to dashboard Cite

We propose a three-dimensional (3D) object recognition approach via computational integral imaging and scale invariant feature transform (SIFT) that can be invariance to object changes in illumination, scale, rotation and affine. Usually, the matching between features extracted in reference object and that in computationally reconstructed image should be done for 3D object recognition. However, this process needs to alternately illustrate all of the depth images first which will affect the recognition efficiency. Considering that there are a set of elemental images with different viewpoint in integral imaging, we first recognize the object in 2D image by using five elemental images and then choose one elemental image with the most matching points from the five images. This selected image will include more information related to the reference object. Finally, we can use this selected elemental image and its neighboring elemental images which should also contain much reference object information to calculate the disparity with SIFT algorithm. Consequently, the depth of the 3D object can be achieved with stereo camera theory and the recognized 3D object can be reconstructed in computational integral imaging. This method sufficiently utilizes the different information provided by elemental images and the robust feature extraction SIFT algorithm to recognize 3D objects.

show abstract

An application of scale-invariant feature transform in iris recognition

Cited by 3 publications

References 13 publications

Snake classification from images

Snake classification from images

Lightweight air-to-air unmanned aerial vehicle target detection model

Three-dimensional object recognition via integral imaging and scale invariant feature transform

Contact Info

Product

Resources

About