Improving SIFT accuracy with use of perspective transforms
نویسنده
چکیده
Feature extraction is a widely used technique in the field of image recognition, multimedia information retrieval. known techniques for this feature extractions are ORB, SIFT, SURF, Wavelets and many more. One part of MIR is object recognition, and with object recognition comes object finding, or homograph finding. This uses a method called RANSAC to determine where an object is in a scene, using descriptors that were generated by for example SIFT. In this project, the focus was on the SIFT method for feature extraction. A former project in the course was shown during the course which had an implementation of SIFT with homograph finding which could track an object and draw a rectangle around it. This actually worked pretty well, except for when the object was turned away from the camera when it quickly started to wobble around since it could not locate the object anymore. This interested me and I decided to try to find a way to improve the accuracy of SIFT when objects were turned away. This would also be interesting to implement in a working example, just like the object tracking example of the project of last year. OpenCV actually has an example project where exactly that is done, but it is not a live example yet.
منابع مشابه
Reduced SIFT Features For Image Retrieval and Indoor Localisation
SIFT features are distinctive invariant features used to robustly describe and match digital image content between different views of a scene. While invariant to scale and rotation, and robust to other image transforms, the SIFT feature description of an image is typically large and slow to compute. This paper presents a method to reduce the size, complexity and matching time of SIFT feature se...
متن کاملCalculation of One-dimensional Forward Modelling of Helicopter-borne Electromagnetic Data and a Sensitivity Matrix Using Fast Hankel Transforms
The helicopter-borne electromagnetic (HEM) frequency-domain exploration method is an airborne electromagnetic (AEM) technique that is widely used for vast and rough areas for resistivity imaging. The vast amount of digitized data flowing from the HEM method requires an efficient and accurate inversion algorithm. Generally, the inverse modelling of HEM data in the first step requires a precise a...
متن کاملCounter-forensics of SIFT-based copy-move detection by means of keypoint classification
Copy-move forgeries are very common image manipulations that are often carried out with malicious intents. Among the techniques devised by the Image Forensic community, those relying on SIFT features are the most effective ones. In this paper we approach the copy-move scenario from the perspective of an attacker whose goal is to remove such features. The attacks conceived so far against SIFT-ba...
متن کاملExamining the Effects of Key Point Detector and Descriptors on 3D Visual SLAM
Mobile robots need to continuously navigate their environment. Doing so necessitates using sensor data to both map that environment and locate their position. A modular framework for performing 3D Simultaneous Localization and Mapping (SLAM) for use with indoor robots has been developed that addresses this problem. This framework was developed using a Microsoft KinectTM sensor and works by extr...
متن کاملReal-Time Scale Invariant Object Recognition Using an Artificial Neural Network
In order to satisfy a need of real-time object identification a scale invariant feature matching method using an artificial neural network (ANN) is proposed. Scale Invariant Feature Transforms (SIFT) feature detection and feature description is compared with the SIFT feature detection and a custom feature descriptor matched with an appropriately trained ANN. The proposed method requires less ti...
متن کامل