- 标题
- 摘要
- 关键词
- 实验方案
- 产品
-
Research on feature point extraction and matching machine learning method based on light field imaging
摘要： At present, there are many methods to realize the matching of specified images with features, and the basic components include image feature point detection, feature description, and image matching. Based on this background, this article has done different research and exploration around these three aspects. The image feature point detection method is firstly studied, which commonly include image edge information-based feature detection method, corner information-based detection method, and various interest operators. However, all of the traditional detection methods are involved in problems of large computation burden and time consumption. In order to solve this problem, a feature detection method based on image grayscale information-FAST operator is used in this paper, which is combined with decision tree theory to effectively improve the speed of extracting image feature points. Then, the feature point description method BRIEF operator is studied, which is a local expression of detected image feature points based on descriptors. Since the descriptor does not have rotation invariance, the detection operator is endowed by a direction that is proposed in this paper, and then the local feature description is conducted on the feature descriptor to generate a binary string array containing direction information. Finally, the feature matching machine learning method is analyzed, and the nearest search method is used to find the nearest feature point pair in Euclidean distance, of which the calculation burden is small. The simulation results show that the proposed nearest neighbor search and matching machine learning algorithm has higher matching accuracy and faster calculation speed compared with the classical feature matching algorithm, which has great advantages in processing a large number of array images captured by the light field camera.
关键词： Nearest neighbor search,Light field imaging,Image matching,Machine learning
更新于2025-09-23 15:23:52
-
[IEEE 2018 26th European Signal Processing Conference (EUSIPCO) - Rome (2018.9.3-2018.9.7)] 2018 26th European Signal Processing Conference (EUSIPCO) - Ear Presentation Attack Detection: Benchmarking Study with First Lenslet Light Field Database
摘要： Ear recognition has received broad attention from the biometric community and its emerging usage in multiple applications is raising new security concerns, with robustness against presentation attacks being a very active field of research. This paper addresses for the first time the ear presentation attack detection problem by developing an exhaustive benchmarking study on the performance of state-of-the-art light field and non-light field based ear presentation attack detection solutions. In this context, this paper also proposes an appropriate ear artefact database captured with a Lytro ILLUM lenslet light field camera, including both 2D and light field contents, using several types of presentation attack instruments, including laptop, tablet and two different mobile phones. Results show very promising performance for two recent light field based presentation attack detection solutions originally proposed for face presentation attack detection.
关键词： Ear Presentation Attack Detection,Feature Extraction,Light Field Imaging,Artefact Database
更新于2025-09-23 15:22:29
-
Light-Field Imaging for Plasma Wind-Tunnel Application
摘要： Limited optical access and high apparatus complexity are the main challenges for true three-dimensional (3-D) measurements in high-enthalpy plasma flows. However, with the advent of plenoptic cameras, the one-shot and single-camera acquisition of light-field data has become possible, enabling the 3-D analysis of complex scenes with one single optical access and one single exposure. So far, this technique has mostly been applied to problems containing opaque objects or particle-loaded flows. In this paper, approaches to explore the potential of light-field analysis for the 3-D investigation of brightly luminous, high-velocity plasma flows are presented. Using gas flames from a solder torch, the feasibility of plenoptic imaging of optically thin scenes is shown. The complete structure of the flame is derived from a single exposure. The transition of this approach to plasma-flow visualization is shown with very first acquisitions of a plasma freejet, including spectral filters for the measurement of the atomic oxygen distribution.
关键词： high-enthalpy flows,light-field imaging,plasma wind tunnel,plenoptic camera,3-D reconstruction
更新于2025-09-23 15:22:29
-
[IEEE 2018 26th European Signal Processing Conference (EUSIPCO) - Rome (2018.9.3-2018.9.7)] 2018 26th European Signal Processing Conference (EUSIPCO) - Road Surface Crack Detection using a Light Field Camera
摘要： During traditional road surveys, inspectors capture images of pavement surface using cameras that produce 2D images, which can then be automatically processed to get a road surface condition assessment. This paper proposes a novel crack detection system that uses a light field imaging sensor, notably the Lytro Illum camera, instead of a conventional 2D camera, to capture road surface light field images. Light field images capture the light rays originating from different directions, thus providing a richer representation of the observed scene. The proposed system explores the disparity information, which can be computed from the light field, to obtain information about cracks observable in the pavement images. A simple processing system is considered, to show the potential use of this type of sensors for crack detection. Encouraging experimental crack detection results are presented based on a set of road pavement light field images captured over different pavement surface textures. A performance comparison with a state-of-the-art 2D image crack detection system is included, confirming the potential of using this type of sensors.
关键词： Light field imaging,image processing,road crack detection
更新于2025-09-19 17:15:36
-
Research on Depth Estimation Method of Light Field Imaging Based on Big Data in Internet of Things From Camera Array
摘要： In recent years, optical field imaging technology has received extensive attention in the academic circle for its novel imaging characteristics of shooting first and focusing later, variable depth of field, variable viewpoint, and so on. However, the existing optical field acquisition equipment can only acquire a limited number of discrete angle signals, so image aliasing caused by under sampling of optical field angle signals reduces the quality of optical field images. Based on the camera array system as a platform, this paper studies the optical field imaging and depth estimation method based on the Big Data in Internet of Things obtained from camera array around the angle sampling characteristics of the optical field data set, and has achieved some innovative research results in the following aspects. On the basis of analyzing the characteristics of different depth clues in the optical field data set, a depth estimation method combining parallax method and focusing method is proposed. First, this paper analyzes the disparity clues and focus clues contained in the multi-view data set and the light field refocusing image set of the camera array, respectively, and points out the differences and relationships between the two depth clues extraction methods in the light field sampling frequency domain space, that is, the disparity method focuses on the energy concentration characteristics near the frequency domain spatial angle axis, while the focus method focuses on the high frequency proportion of energy distribution on the angle axis. Then, the weighted linear fusion method based on image gradient is used to fuse the two calculation results, which improves the accuracy and robustness of depth estimation. Finally, the results of depth estimation experiments on different sets of scenes show that compared with the method based on a single depth cue, the method in this paper has higher accuracy in depth calculation in discontinuous areas of scene depth and similar texture areas.
关键词： light field imaging,camera array,depth estimation,image aliasing,Big data in Internet of Things(IoT)
更新于2025-09-10 09:29:36
-
[IEEE 2018 IEEE 28th International Workshop on Machine Learning for Signal Processing (MLSP) - Aalborg, Denmark (2018.9.17-2018.9.20)] 2018 IEEE 28th International Workshop on Machine Learning for Signal Processing (MLSP) - LIGHT FIELD BASED FACE RECOGNITION VIA A FUSED DEEP REPRESENTATION
摘要： The emergence of light field cameras opens new frontiers in terms of biometric recognition. This paper proposes the first deep CNN solution for light field based face recognition, exploiting the richer information available in a lenslet light field image. Additionally, for the first time, the exploitation of disparity maps together with 2D-RGB images and depth maps has been considered in the context of the face recognition a fusion scheme performance. The proposed solution uses the 2D-RGB central sub-aperture view as well as the disparity and depth maps extracted from the full set of sub-aperture images associated to a lenslet light field. After, feature extraction is performed using a VGG-Face deep descriptor for texture and independently fine-tuned models for disparity and depth maps. Finally, the extracted features are concatenated to be fed into an SVM classifier. A comprehensive set of experiments has been conducted with the IST-EURECOM light field face database, showing the superior performance of the fused deep representation for varied and challenging recognition tasks.
关键词： Disparity Map,Face Recognition,Depth Map,Lenslet Light Field Imaging,Fine-Tuning,VGG-Face Descriptor
更新于2025-09-09 09:28:46