Abstract This paper presents a face recognition system that is part of a global solution for online soccer piracy detection. The overall solution uses several building blocks to detect illegal sharing of live soccer broadcasts. This paper presents one of the building blocks, a face recognition system, that recognizes the faces of players that participate in the soccer match that the overall solution is trying to identify. The face recognition system detects faces in the broadcast image frames and tries to match them to a database faces of players from the clubs that participate in the game. The proposed face recognition system uses Retinaface and OpenCV to detect faces, applies Deep Learning networks Facenet128 and FaceNet512 to extract features from the detected faces, computes the cosine distance between features to evaluate the dissimilarity between faces, and compares the distance to a predefined threshold. This approach aims to maximize Precision and True Positive Rate while ensuring a False Positive Rate equal to zero, even at the cost of a lower Recall and Accuracy, and also provides results in as close to real-time as possible. The experiments show that the proposed face recognition system is able to achieve a True Positive Rate of 38.4% while ensuring a False Positive Rate of 0, which is an important aspect for the overall solution. The system is able to analyze an average of 20 frames per second. The results show the potential of this approach to identify and combat illegal broadcasts of sporting events, offering a robust approach to address the escalating issue of unauthorized audiovisual content sharing.

In cameras with radial distortion, straight lines in space are in general mapped to curves in the image. Although epipolar geometry also gets distorted, there is a set of special epipolar lines that remain straight, namely those that go through the distortion center. By finding these straight epipolar lines in camera pairs we can obtain constraints on the distortion center(s) without any calibration object or plumb line assumptions in the scene. Although this holds for all radial distortion models we conceptually prove this idea using the division distortion model and the radial fundamental matrix which allow for a very simple closed form solution of the distortion center from two views (same distortion) or three views (different distortions). The non-iterative nature of our approach makes it immune to local minima and allows finding the distortion center also for cropped images or those where no good prior exists. Besides this, we give comprehensive relations between different undistortion models and discuss advantages and drawbacks.

For modern consumer cameras often approximate calibration data is available, mak-ing applications such as 3D reconstruction or photo registration easier as compared to the pure uncalibrated setting. In this paper we address the setting with calibrated-uncalibrated image pairs: for one image intrinsic parameters are assumed to be known, whereas the second view has unknown distortion and calibration parameters. This sit-uation arises e.g. when one would like to register archive imagery to recently taken photos. A commonly adopted strategy for determining epipolar geometry is based on feature matching and minimal solvers inside a RANSAC framework. However, only very few existing solutions apply to the calibrated-uncalibrated setting. We propose a simple and numerically stable two-step scheme to first estimate radial distortion parameters and subsequently the focal length using novel solvers. We demonstrate the performance on synthetic and real datasets. 1