3D spatial recognition is a fundamental technology that supports automatic driving. For example, the processing accuracy of the vehicle depends on the accuracy of depth information around the vehicle body. While methods to geometrically measure the depth of the captured space by applying stereo vision to images taken by multiple cameras become being widely used, it is difficult to measure depth in poorly textured or occluded regions. On the other hand, it becomes possible to estimate depth information in such areas with the advent of estimating depth from monocular images by deep learning. However, if the observation conditions differ between training and estimation, the accuracy of the estimation will decline. This paper proposes a complementary method that integrates both methods by using a convolutional autoencoder.
Catcher framing technique in baseball has recently gained attention in game analysis. This technique involves a catcher adjusting their catching motion to increase the likelihood of an umpire calling a pitch a strike. Its success is typically evaluated based on the strike rate at the boundary of the strike zone, calculated from pitch trajectory data obtained through a tracking system. However, evaluating catcher framing in games without a tracking system is challenging, and alternative methods based on different types of information are needed. This research proposes a method to detect the catcher's mitt movement trajectory during catcher framing, which is considered useful information apart from pitch trajectory. The method applies object detection, pose estimation and deep learning to videos of baseball pitching scenes.
Camera work is a critical aspect of conveying the atmosphere and impressions in movies, and it plays a vital role in video analysis. This research proposes a method to estimate camera work from monocular videos by analyzing the optical flow within the video frames. Our method facilitates the estimation of camera work in videos featuring dynamic subjects by incorporating semantic segmentation. Additionally, it is capable of distinguishing between zoom and dolly movements, which previous works have not achieved. The method uses the relationship between image depth, optical flow, and image coordinates to perform such classification.
Micro tunnel excavation is limited to straight paths due to size constraints that prevent human entry through the jacking pipes. A compact, wide-angle system enables the measurement of enclosed spaces with complex paths. This paper proposes a 3D estimation method using a catadioptric imaging system, which consists of an omnidirectional camera and a single spherical mirror, to capture monocular images without relying on training data. The applicability of our method is demonstrated by measuring the geometry of a pipe designed to imitate a propulsion pipe used in tunnel excavation.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.