Free-Viewpoint Video for TV Sport Production

Hilton, Adrian; Guillemaut, Jean-Yves; Kilner, J.; Grau, Oliver; Thomas, Graham

doi:10.1007/978-3-642-12392-4_4

Cited by 6 publications

(1 citation statement)

References 49 publications

(66 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Computer vision has been widely used in many sports applications [43]. The number of applications has expanded from information extraction such as player detection and tracking [38] to new visual information generation such as free-viewpoint video generation [25], and further to prediction of shot location [63] and broadcast camera angle planning [12].…”

Section: Introductionmentioning

confidence: 99%

Classification of Puck Possession Events in Ice Hockey

Tora

Chen

Little

2017

2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)

View full text Add to dashboard Cite

Group activity recognition in sports is often challenging due to the complex dynamics and interaction among the players. In this thesis, we propose a deep architecture to classify puck possession events in ice hockey. Our model consists of three distinct phases: feature extraction, feature aggregation and, learning and inference. For the feature extraction and aggregation, we use a Convolutional Neural Network (CNN) followed by a late fusion model on top to extract and aggregate different types of features that includes handcrafted homography features for encoding the camera information. The output from the CNN is then passed into a Recurrent Neural Network (RNN) for the temporal extension and classification of the events. The proposed model captures the context information from the frame features as well as the homography features. The individual attributes of the players and the interaction among them is also incorporated using a pre-trained model and team pooling. Our model requires only the player positions on the image and the homography matrix and does not need any explicit annotations for the individual actions or player trajectories, greatly simplifying the input of the system. We evaluate our model on a new Ice Hockey Dataset and a Volleyball Dataset. Experimental results show that our model produces promising results on both these challenging datasets with much simpler inputs compared with the previous work. ii Lay Summary Group activity recognition is the task of determining what a group of people are doing given a single image or a short clip of video. We have looked at group activity recognition in sports videos, particularly ice hockey. Thus given a sequence of images, we aim to classify the sequence into a group activity or event. There are many possible events that can happen in ice hockey but we have looked at a subset of only those events which involve the possession of the puck by the players. Some examples include pass and shot. We have solved this problem by proposing a deep network architecture which takes into account player appearance and contextual information. These features from different sources are fused together and passed into a temporal model to learn the dependencies across the images in the given sequence. iii Preface This thesis is submitted in partial fulfillment of the requirements for a Master of Science Degree in Computer Science. The entire work presented here is original work done by the author, Moumita Roy Tora, performed under the supervision of Professor James J. Little.

show abstract

Section: Introductionmentioning

confidence: 99%

Classification of Puck Possession Events in Ice Hockey

Tora

Chen

Little

2017

2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)

View full text Add to dashboard Cite

show abstract

Immersion, Präsenzerleben und Transportation. 360-Grad-Video in der Sportkommunikation

Hebbel-Seeger

2024

Sport in Audiovisuellen Medien

View full text Add to dashboard Cite

Smooth switching method for asynchronous multiple viewpoint videos using frame interpolation

Shishido

Harazaki

Kameda

et al. 2019

Journal of Visual Communication and Image Representation

View full text Add to dashboard Cite

This research proposes a method that generates viewpoint smooth switching by reducing the flickering artefacts that are observed at bullet-times generated from asynchronous multi-view videos using frame interpolation processing. When we asynchronously capture multi-view videos of an object moving at a high velocity, deviations occur in the observed position at the bullet-times. We apply a frame interpolation technique to smooth this problem. By selecting suitable interpolated images that produce the smallest movement of the subject's observed position, we smoothly generate a viewpoint-switched bullet-time video. In this paper, we examine the subjective evaluation of the video generated by the proposed method. And we also examine objective evaluation. Therefore, the effectiveness of the proposed method is shown. Furthermore, reproducibility was improved by considering the application conditions of the proposed method.

show abstract

Free-Viewpoint Video for TV Sport Production

Cited by 6 publications

References 49 publications

Classification of Puck Possession Events in Ice Hockey

Classification of Puck Possession Events in Ice Hockey

Immersion, Präsenzerleben und Transportation. 360-Grad-Video in der Sportkommunikation

Smooth switching method for asynchronous multiple viewpoint videos using frame interpolation

Contact Info

Product

Resources

About