Computer Vision Projects
An international project focused on building the best embedded computer vision platform ever. EoT is an Innovation Action funded under the European Union’s H2020 Framework Programme that is going to unfold for three years starting in January 2015.Eyes of Things
Its not complete yet, its only tracking points and it hasnt got a memory yet.
Theres a few more videos on my youtube about it. rouncer81 channel.
CANDELOR is a computer vision library for 3D scene interpretation. It sports high performance object localization algorithms that are robust and simple to use.CANDELOR
The rotation between the camera and the environment is found from the directions of edgels extracted from the image. Any camera model / projection can be used. This example is an equirectangular image, from Google StreetView.Corisco - camera orientation from infinitesimal segments by constrained optimization
The goal of this project is to land a UAV on a runway using computer vision. The UAV has no sensors other than a forward-looking video camera. The video is transmitted in real time to a ground-station that uses image registration to detect and determine the orientation of the runway in each frame. The ground-station sends radio commands back to the UAV to guide it to the runway.Landing a UAV on a Runway Using Image Registration
A team of University of Notre Dame biometrics experts is developing a crime-fighting tool that can help law enforcement and military officials identify suspicious individuals at crime scenes.Questionable Observer Detector Biometrics System
The goal of this project was to elaborate a computer vision based method for the automatic tracking of golf clubs. It can provide additional information to the users such as club trajectory, speed and acceleration.Robust Visual Golf Club Tracking
Tracking multiple people from standard cameras is challenging, mostly due to the occlusions that occur as soon as several people are involved. We tackle this problem by using several cameras, observing the scene with different points of views. We developed a people detection algorithm called POM that uses a generative model of background subtraction to estimate the positions of people in an individual time frame.Tracking People using Multiple Cameras
Detecting and reading text from natural images is a hard computer vision task that is central to a variety of emerging applications. Related problems like document character recognition have been widely studied by computer vision and machine learning researchers and are virtually solved for practical applications like reading handwritten digits. Reliably recognizing characters in more complex scenes like photographs, however, is far more difﬁcult: the best existing methods lag well behind human performance on the same tasks. In this paper we attack the problem of recognizing digits in a real application using unsupervised feature learning methods: reading house numbers from street level photos.Reading Digits in Natural Images with Unsupervised Feature Learning
This paper presents a system for performance-based character animation that enables any user to control the facial expressions of a digital avatar in realtime. The user is recorded in a natural environment using a non-intrusive, commercially available 3D sensor. The simplicity of this acquisition device comes at the cost of high noise levels in the acquired data. To effectively map low-quality 2D images and 3D depth maps to realistic facial expressions, we introduce a novel face tracking algorithm that combines geometry and texture registration with pre-recorded animation priors in a single optimization.Realtime Performance-Based Facial Animation (SIGGRAPH 2011)
Panoramic photography creates fascinating images. Very wide angle images are closer to the human field of view than conventional pictures. If seen through a panoramic viewer they let us experience a location as if we were there. Panoramic image stitching can create panoramas from pictures taken one after another. Unfortunately, acquiring the images takes a lot of time and moving objects may cause ghosting. It is also difficult to obtain a full spherical panorama, because the downward picture cannot be captured while the camera is mounted on the tripod.Throwable Panoramic Ball Camera
Weakly supervised discovery of correspondence among a set of complex, cluttered images is one of the key problems in recognition. We address this problem using deformable part-based models (DPM's) with latent SVM training. This method has originally been developed for fully supervised training of object detectors, but we demonstrate that it is also capable of more open-ended discovery of latent common structure for tasks such as scene recognition and weakly-supervised object localization.Scene Recognition and Weakly Supervised Object Localization with Deformable Part-Based Models
The goal of this work is to find visually similar images even if they appear quite different at the raw pixel level. This task is particularly important for matching images across visual domains, such as photos taken over different seasons or lighting conditions, paintings, hand-drawn sketches, etc. We propose a surprisingly simple method that estimates the relative importance of different features in a query image based on the notion of "data-driven uniqueness".Data-driven Visual Similarity for Cross-domain Image Matching