Showing papers on "Zoom published in 2017"

Adaptive foveated single-pixel imaging with dynamic supersampling

Abstract: In contrast to conventional multipixel cameras, single-pixel cameras capture images using a single detector that measures the correlations between the scene and a set of patterns. However, these systems typically exhibit low frame rates, because to fully sample a scene in this way requires at least the same number of correlation measurements as the number of pixels in the reconstructed image. To mitigate this, a range of compressive sensing techniques have been developed which use a priori knowledge to reconstruct images from an undersampled measurement set. Here, we take a different approach and adopt a strategy inspired by the foveated vision found in the animal kingdom—a framework that exploits the spatiotemporal redundancy of many dynamic scenes. In our system, a high-resolution foveal region tracks motion within the scene, yet unlike a simple zoom, every frame delivers new spatial information from across the entire field of view. This strategy rapidly records the detail of quickly changing features in the scene while simultaneously accumulating detail of more slowly evolving regions over several consecutive frames. This architecture provides video streams in which both the resolution and exposure time spatially vary and adapt dynamically in response to the evolution of the scene. The degree of local frame rate enhancement is scene-dependent, but here, we demonstrate a factor of 4, thereby helping to mitigate one of the main drawbacks of single-pixel imaging techniques. The methods described here complement existing compressive sensing approaches and may be applied to enhance computational imagers that rely on sequential correlation measurements.

Gaze + pinch interaction in virtual reality

Abstract: Virtual reality affords experimentation with human abilities beyond what's possible in the real world, toward novel senses of interaction. In many interactions, the eyes naturally point at objects of interest while the hands skilfully manipulate in 3D space. We explore a particular combination for virtual reality, the Gaze + Pinch interaction technique. It integrates eye gaze to select targets, and indirect freehand gestures to manipulate them. This keeps the gesture use intuitive like direct physical manipulation, but the gesture's effect can be applied to any object the user looks at --- whether located near or far. In this paper, we describe novel interaction concepts and an experimental system prototype that bring together interaction technique variants, menu interfaces, and applications into one unified virtual experience. Proof-of-concept application examples were developed and informally tested, such as 3D manipulation, scene navigation, and image zooming, illustrating a range of advanced interaction capabilities on targets at any distance, without relying on extra controller devices.

CNN based semantic segmentation for urban traffic scenes using fisheye camera

Abstract: Semantic segmentation is an important step of visual scene understanding for autonomous driving. Recently, Convolutional Neural Network (CNN) based methods have successfully applied in semantic segmentation using narrow-angle or even wide-angle pinhole camera. However, in urban traffic environments, autonomous vehicles need wider field of view to perceive surrounding things and stuff, especially at intersections. This paper describes a CNN-based semantic segmentation solution using fisheye camera which covers a large field of view. To handle the complex scene in the fisheye image, Overlapping Pyramid Pooling (OPP) module is proposed to explore local, global and pyramid local region context information. Based on the OPP module, a network structure called OPP-net is proposed for semantic segmentation. The net is trained and evaluated on a fisheye image dataset for semantic segmentation which is generated from an existing dataset of urban traffic scenes. In addition, zoom augmentation, a novel data augmentation policy specially designed for fisheye image, is proposed to improve the net's generalization performance. Experiments demonstrate the outstanding performance of the OPP-net for urban traffic scenes and the effectiveness of the zoom augmentation.

Dual field-of-view step-zoom metalens.

Abstract: A conventional optical zoom system is bulky, expensive, and complicated for real-time adjustment. Recent progress in metasurface research has provided a new solution to achieve innovative compact optical systems. In this Letter, we propose a highly integrated step-zoom lens with dual field of view (FOV) based on double-sided metasurfaces. With silicon nanobrick arrays of spatially varying orientations sitting on both sides of a transparent substrate, this ultrathin step-zoom metalens can be designed to focus an incident circular polarized beam with handedness-dependent FOVs without varying the focal plane, which is important for practical applications. The proposed dual FOV step-zoom metalens, with advantages such as ultracompactness, flexibility, and replicability, can find applications in fields that require ultracompact zoom imaging and beam focusing.

Zoom out CNNs features for optical remote sensing change detection

Abstract: In this paper, we propose a novel unsupervised optical remote sensing change detection (CD) based on pre-trained convolutional neural network (CNN) on ImageNet dataset and superpixel (SLIC) segmentation technique. The proposed approach can be divided into three steps. First, bi-temporal images are stacked, and Principal Component Analysis (PCA) is applied to extract three higher uncorrelated channels, which will be later segmented into superpixels. Second, we zoom out each region into three levels and fit them separately into a pre-trained CNN. Third, we extract features of different zooming levels that represent the same region (superpixel) and concatenate them. We compare the concatenated features to get the final change map. The experimental results demonstrate the efficacy of the proposed approach.

HMD transitions for focusing on specific content in virtual-reality environments

Abstract: Methods and systems for presenting an object on to a screen of a head mounted display (HMD) include receiving an image of a real-world environment in proximity of a user wearing the HMD. The image is received from one or more forward facing cameras of the HMD and processed for rendering on a screen of the HMD by a processor within the HMD. A gaze direction of the user wearing the HMD, is detected using one or more gaze detecting cameras of the HMD that are directed toward one or each eye of the user. Images captured by the forward facing cameras are analyzed to identify an object captured in the real-world environment that is in line with the gaze direction of the user, wherein the image of the object is rendered at a first virtual distance that causes the object to appear out-of-focus when presented to the user. A signal is generated to adjust a zoom factor for lens of the one or more forward facing cameras so as to cause the object to be brought into focus. The adjustment of the zoom factor causes the image of the object to be presented on the screen of the HMD at a second virtual distance that allows the object to be discernible by the user.

All-liquid dual-lens optofluidic zoom system

Abstract: An all-liquid optical zoom system based on two independently controllable liquid lenses positioned inside a single sealed cylindrical housing is demonstrated. This system yielded a zoom ratio of 1.5 for an object distance of 200 mm, an image distance 37 mm, and a corresponding resolution of better than 5 line pairs/mm. With a diameter of 5 mm, a packaged system length of 9.88 mm, and a power consumption of 3.5 mW, the system represents a new generation of ultra-miniaturized optofluidic systems with high functionality, excellent imaging properties, and highly flexible tunability, all with no mechanically moving parts.

Zoom Out-and-In Network with Recursive Training for Object Proposal

Abstract: In this paper, we propose a zoom-out-and-in network for generating object proposals. We utilize different resolutions of feature maps in the network to detect object instances of various sizes. Specifically, we divide the anchor candidates into three clusters based on the scale size and place them on feature maps of distinct strides to detect small, medium and large objects, respectively. Deeper feature maps contain region-level semantics which can help shallow counterparts to identify small objects. Therefore we design a zoom-in sub-network to increase the resolution of high level features via a deconvolution operation. The high-level features with high resolution are then combined and merged with low-level features to detect objects. Furthermore, we devise a recursive training pipeline to consecutively regress region proposals at the training stage in order to match the iterative regression at the testing stage. We demonstrate the effectiveness of the proposed method on ILSVRC DET and MS COCO datasets, where our algorithm performs better than the state-of-the-arts in various evaluation metrics. It also increases average precision by around 2% in the detection system.

System and method for locating and performing fine grained classification from multi-view image data

Abstract: Some embodiments of the invention provide a method for identifying geographic locations and for performing a fine-grained classification of elements detected in images captured from multiple different viewpoints or perspectives. In several embodiments, the method identifies the geographic locations by probabilistically combining predictions from the different viewpoints by warping their outputs to a common geographic coordinate frame. The method of certain embodiments performs the fine-grained classification based on image portions from several images associated with a particular geographic location, where the images are captured from different perspectives and/or zoom levels.

Ultra-compact optical zoom endoscope using solid tunable lenses.

Abstract: We report an ultra-compact optical zoom endoscope containing two tunable Alvarez lenses. The two tunable lenses are controlled synchronously by piezoelectric benders to move in directions perpendicular to the optical axis to achieve optical zoom while keeping images in clear focus without moving the scope. The piezoelectric benders are arranged circumferentially surrounding the endoscope optics with a diameter about 2 mm, which results in an ultra-compact form. The demonstrated endoscope is capable of optical zoom close to 3 × from field of view (FOV) 50° to 18° continuously with the required movements for its constituent optical elements less than 110 μm. Such optical zoom endoscopes may find their potential uses in healthcare and industrial inspection systems.

Terminal and controlling method thereof

Abstract: Disclosed are a terminal and controlling method thereof. The present invention includes displaying a preview image captured through a camera, obtaining an input for selecting an object from the displayed preview image, determining a smart zoom magnification for the selected object to be displayed in an optimal size, zoom-processing the preview image at the determined smart zoom magnification centering on the selected object, and displaying the zoom-processed preview image.

Panning and Zooming High-Resolution Panoramas in Virtual Reality Devices

Abstract: Two recent innovations in immersive media include the ability to capture very high resolution panoramic imagery, and the rise of consumer level heads-up displays for virtual reality. Unfortunately, zooming to examine the high resolution in VR breaks the basic contract with the user, that the FOV of the visual field matches the FOV of the imagery. In this paper, we study methods to overcome this restriction to allow high resolution panoramic imagery to be able to be explored in VR. We introduce and test new interface modalities for exploring high resolution panoramic imagery in VR. In particular, we demonstrate that limiting the visual FOV of the zoomed in imagery to the central portion of the visual field, and modulating the transparency or zoom level of the imagery during rapid panning, reduce simulator sickness and help with targeting tasks.

Magnifying Smartphone Screen Using Google Glass for Low-Vision Users

Abstract: Magnification is a key accessibility feature used by low-vision smartphone users. However, small screen size can lead to loss of context and make interaction with magnified displays challenging. We hypothesize that controlling the viewport with head motion can be natural and help in gaining access to magnified displays. We implement this idea using a Google Glass that displays the magnified smartphone screenshots received in real time via Bluetooth. Instead of navigating with touch gestures on the magnified smartphone display, the users can view different screen locations by rotating their head, and remotely interacting with the smartphone. It is equivalent to looking at a large virtual image through a head contingent viewing port, in this case, the Glass display with ~ 15 ° field of view. The system can transfer seven screenshots per second at 8 × magnification, sufficient for tasks where the display content does not change rapidly. A pilot evaluation of this approach was conducted with eight normally sighted and four visually impaired subjects performing assigned tasks using calculator and music player apps. Results showed that performance in the calculation task was faster with the Glass than with the phone's built-in screen zoom. We conclude that head contingent scanning control can be beneficial in navigating magnified small smartphone displays, at least for tasks involving familiar content layout.

System and method for outputting a haptic effect based on a camera zoom state, camera perspective, and/or a direction in which a user's eyes are directed

Abstract: A haptic peripheral system comprising a display, a haptic peripheral device, and a processor is disclosed. The processor is in communication with the display and the haptic peripheral device. The processor is configured to determine a zoom state of a virtual camera of a virtual environment, and to determine at least one of: (i) a perspective of the virtual camera, wherein the perspective indicates which of a first person point of view and a third person point of view the virtual camera has, and (ii) a direction in which the user's eyes are directed. The processor is further configured to cause the haptic peripheral device to output a haptic effect that depends on the zoom state, and on at least one of: (i) the perspective of the virtual camera and (ii) the direction in which a user's eyes are directed.

Electrically optofluidic zoom system with a large zoom range and high-resolution image.

Abstract: We report an electrically controlled optofluidic zoom system which can achieve a large continuous zoom change and high-resolution image. The zoom system consists of an optofluidic zoom objective and a switchable light path which are controlled by two liquid optical shutters. The proposed zoom system can achieve a large tunable focal length range from 36mm to 92mm. And in this tuning range, the zoom system can correct aberrations dynamically, thus the image resolution is high. Due to large zoom range, the proposed imaging system incorporates both camera configuration and telescope configuration into one system. In addition, the whole system is electrically controlled by three electrowetting liquid lenses and two liquid optical shutters, therefore, the proposed system is very compact and free of mechanical moving parts. The proposed zoom system has potential to take place of conventional zoom systems.

Survey on zoom-lens calibration methods and techniques

Abstract: This paper surveys zoom-lens calibration approaches, such as pattern-based calibration, self-calibration, and hybrid (or semiautomatic) calibration. We describe the characteristics and applications of various calibration methods employed in zoom-lens calibration and offer a novel classification model for zoom-lens calibration approaches in both single and stereo cameras. We elaborate on these calibration techniques to discuss their common characteristics and attributes. Finally, we present a comparative analysis of zoom-lens calibration approaches, highlighting the advantages and disadvantages of each approach. Furthermore, we compare the linear and nonlinear camera models proposed for zoom-lens calibration and enlist the different techniques used to model the camera’s parameters for zoom (or focus) settings.

Designing computer-based learning contents: influence of digital zoom on attention

Abstract: In the present study, we investigated the role of digital zoom as a tool for directing attention while looking at visual learning material In particular, we analyzed whether minimal digital zoom functions similarly to a rhetorical device by cueing mental zooming of attention accordingly Participants were presented either static film clips, film clips with minimal zoom-ins, or film clips with minimal zoom-outs while eye movements were recorded We hypothesized that minimal zoom-ins should lead to more gaze coherence, to longer dwell times as an indicator of more elaborative processing, and to fewer transitions as an indicator of less mental integration Zoom-outs, on the other hand, were expected to have opposite effects Results showed that zoom-ins increase gaze coherence and dwell times on the center parts of the depictions while decreasing transitions of pictorial elements from the center and the context areas In contrast, patterns of results from zoom-outs and static presentations were similar to a large degree, indicating that zoom-ins and zoom-outs do not operate in a complementary fashion Theoretical and practical implications of the present results are discussed

Methods and systems for auto-zoom based adaptive video streaming

Abstract: Systems, methods, and computer readable media are described for providing automatic zoom based adaptive video streaming. In some examples, a tracking video stream and a target video stream are obtained and are processed. The tracking video stream has a first resolution, and the target video stream has a second resolution that is higher than the first resolution. The tracking video stream is processed to define regions of interest for frames of the tracking video stream. The target video stream is processed to generate zoomed-in regions of frames of the target video stream. A zoomed-in region of the target video stream corresponds to a region of interest defined using the tracking video stream. The zoomed-in regions of the frames of the target video stream are then provided for display on a client device.

Towards an Improved Vision-Based Web Page Segmentation Algorithm

Abstract: In this paper we introduce an edge-based segmentation algorithm designed for web pages. We consider each web page as an image and perform segmentation as the initial stage of a planned parsing system that will also include region classification. The motivation for our work is to enable improved online experiences for users with assistive needs (serving as the back-end process for such front-end tasks as zooming and decluttering the image being presented to those with visual or cognitive challenges, or producing less unwieldy output from screenreaders). Our focus is therefore on the interpretation of a class of man-made images (where web pages consist of one particular set of these images which have important constraints that assist in performing the processing). After clarifying some comparisons with an earlier model of ours, we show validation for our method. Following this, we briefly discuss the contribution for the field of computer vision, offering a contrast with current work in segmentation focused on the processing of natural images.

Automated scheduling of radar-cued camera system for optimizing visual inspection and detection of radar targets

Abstract: A wide field-of-view and rapid response to threats are critical components of any surveillance system. Field of view is normally implemented by articulating a camera allowing it to swivel to pan and tilt, and actively zooming in on “interesting” locations. Since a single camera suffers from the “soda straw” problem, where only a small portion of the scene can be examined at any given time (leaving the rest of the scene unwatched), surveillance systems often employ a radar unit to direct the operator to likely targets. This provides direction to the search, but still poses a security risk, since potentially hazardous activities might be occurring in an unwatched portion of the field of view while the operator is investigating another incident (which can be either coincidentally or intentionally distracting). Today's systems all rely on a human operator to control the slewing of the camera to inspect the potential targets found by the radar. Automated schedulers have thus far been avoided by these systems, since it has always been assumed that the human would outperform the algorithm. This paper describes a method for automatic scheduling and control for a single or multi-camera radar-cued surveillance system that optimizes visual coverage and inspection of radar-detected-targets. The scheduling algorithm combines track life, track spatial density, and the camera slew angle and speed into a single metric to determine next slew and zoom of camera that maximizes the visual detection of all radar hits over a given period of time and can be run in real-time on a laptop or embedded hardware. The goal of this work is to enable the operator to visually inspect as many radar hits as possible over the course of the operator's shift.

Switchover control techniques for dual-sensor camera system

Abstract: Techniques are described for automatically selecting between multiple image capture subsystems with overlapping fields of view but different optical properties. A selection may be made by estimating a plurality of operational characteristics of an image capture event, and, based on those estimates, selecting a primary image capture subsystem for the image capture event. For example, in a device such as a cellphone comprising two capture subsystems, each subsystem including a lens system and sensor system where each subsystem has a different fixed optical zoom parameter, a subsystem can be chosen based on a combination of desired zoom value, estimated focus distance, and estimated scene brightness.

Some Dynamic Aspects of Photogrammetry Missions Performed by “PW-ZOOM” – The UAV of Warsaw University of Technology

Abstract: The article presents the analyses of the flights carried out the by the Unmanned Aerial Vehicle (UAV) named PW-ZOOM used to perform a photogrammetric mission and monitoring of fauna in Antarctic areas. The analyses focus on the deviations of the optical axis of the photo-camera which occurred during photogrammetric flights carried out on the same route but during several Antarctic expeditions performed in subsequent years (2014 and 2015). The results were subjected to correlation tests with weather conditions (wind speed and variability). The basis for these analyses are the data from the onboard signal recorder integrated with an autopilot.

Surveillance automatic tracking method and device

Abstract: The invention provides a surveillance automatic tracking method and device. The method comprises the following steps: monitoring whether a moving object is existent in a monitoring target region when performing the video surveillance; if the moving object is existent in the monitoring target region, drawing the movement trajectory of the moving object, producing a movement control instruction according to the movement trajectory, analyzing the size of the moving object, and producing a zoom control instruction according to the size; adjusting a rotation angle of a speed dome through the movement control instruction, and regulating the focal length of a camera lens according to the zoom control instruction so as to zoom in or zoom out the moving object. By use of the automatic tracking method provided by the invention, the automatic tracking on the target object can be realized, and the picture information of the moving object can be acquired more clearly when the shot moving object is focused through the automatic focusing of the equipment.

Experience over time: evaluating the experience of use of a squeezable interface in the medium term

Abstract: The paper presents the user experience evaluation of Squeeze Me, an interactive cover for tablet and smartphone that enables continuous and expressive interaction with electronic devices. The cover has been used to implement "Squeeze to zoom", a mobile application to zoom in and out while taking a photograph from a tablet. The experience of use was evaluated in the short and medium term, comparing the Squeeze Me interaction modality with classic modalities for zooming in and out commonly available on tablets and smartphones. The evaluation process was conducted using AttrakDiff [3] a questionnaire that measures hedonic stimulation and identity, as well as pragmatic qualities and attractiveness of software products. Participants were asked to try out different interaction modalities for comparison in the short-term (67 people) and over 4 weeks (8 people). Results obtained in the short-term evaluation reveal that "Squeeze to zoom" was awarded higher values than the classic "Slide to zoom" in the hedonic quality-stimulation and attractiveness dimensions, whilst it obtained lower values in the pragmatic quality and hedonic quality-identity. However, the experience of use changed over time. During the longitudinal study, the usability of "Squeeze to zoom" improved whilst the attractiveness of "Slide to zoom" decreases significantly. Furthermore results reveal that "Squeeze to zoom" is significantly more appreciated for its hedonic qualities and the effect is maintained over time. This study highlights the importance of evaluating the experience of use over time, a practice that is almost ignored in the literature on Experience Design.

Automatic chase after burnt device and terminal

Abstract: The utility model provides an automatic chase after burnt device and terminal detection module for detect the motion target in the scene, select a motion target as the target area, use external rectangle frame flag target, ranging module is used for measuring the motion target is worth with the distance of camera, the module zooms for calculate target area degree of depth mean value, and seek and zoom the multiple table, acquire and zoom the multiple, realize zooming, trail the module for the target is trailed the target in next two field picture, and is given result feedback as the input of target following algorithm in the rectangle frame after will zooming it realizes zooming to zoom the module, detection module, ranging module, realize signal connection between zooming the module and trailing the module. Thereby the realization chases after burnt function in real time automatically and bigger focus scope zooms, when target distance is far away, can zoom the target and focus through the exact, make the target clear to keep the suitable size of settlement unchangeable.

Computational zoom: a framework for post-capture image composition

Abstract: Capturing a picture that "tells a story" requires the ability to create the right composition. The two most important parameters controlling composition are the camera position and the focal length of the lens. The traditional paradigm is for a photographer to mentally visualize the desired picture, select the capture parameters to produce it, and finally take the photograph, thus committing to a particular composition. We propose to change this paradigm. To do this, we introduce computational zoom, a framework that allows a photographer to manipulate several aspects of composition in post-processing from a stack of pictures captured at different distances from the scene. We further define a multi-perspective camera model that can generate compositions that are not physically attainable, thus extending the photographer's control over factors such as the relative size of objects at different depths and the sense of depth of the picture. We show several applications and results of the proposed computational zoom framework.

Camera switching method and apparatus, and terminal

Abstract: The invention provides a camera switching method and apparatus, and a terminal. The method comprises the following steps: when it is determined that the current zoom satisfies a camera switching condition, firstly obtaining a distance between a shooting object and a camera and/or current illuminance, then judging whether the distance between the shooting object and the camera and/or the current illuminance satisfies the switching condition, and when the switching condition is satisfies, controlling the camera to switch. Therefore, the camera switching control is realized according to the zoom, the distance and/or the current illuminance, thereby improving the zoom success rate after the camera switching, reducing the number of noise points in a shot image, improving the effect of the shot image, improving the camera performance of the terminal and improving the user experience.

A 3D interaction technique for selection and manipulation distant objects in augmented reality

Abstract: 3D object selection and manipulation is one of the essential features for any augmented reality (AR) system. However, distant object selection and manipulation still suffer from lack of accuracy and precision. This paper introduces an alternate 3D interaction technique for selection and manipulation distant 3D object in in immersive video see-through AR. The proposed interaction technique offers a high precision when selecting and manipulation distant objects thanks to the zooming-based idea. This later, allows bringing closer both real and virtual objects during maintaining the spatio-temporal registration between the virtual and the real scenes. The evaluation of our proposed approach and the comparison with other well-known techniques are given at the end of this paper.