Machine Graphics & Vision

Radar image processing application based on space cloud computing in basketball game guidance camera

2025-05-19T21:27:29+00:00

Capturing and presenting exciting moments is crucial for the audience's experience in basketball game broadcast cameras. However, traditional radar image processing techniques are limited by various factors and cannot meet the demands of modern audiences for high quality, multi angle, and real-time performance. In response to these challenges, an innovative radar image processing system based on space cloud computing has been proposed. Compared with traditional radar image processing systems, the system proposed by the research institute had the best performance, with accuracy, recall, and F1 value reaching 97.08%, 96.88%, and 97.11%, respectively, and a transmission time of only 2.2 seconds; and the stability was greater than 90%, which was about 10% to 25% higher than other systems. In summary, the system proposed by the research institute has brought revolutionary improvements to basketball game guidance and filming through its efficient processing capabilities, accurate image recognition, fast data processing and transmission, and excellent stability. This not only greatly enriches the audience's viewing experience, but also opens up new directions for the development of sports event broadcasting technology. With the further maturity of technology and the continuous expansion of applications, it is expected that this system will play a more important role in future sports event broadcasting, promoting the development of the entire industry towards higher quality and efficiency.

Simple derivation of the Hermite bicubic patch using tensor product

2025-06-15T00:05:32+00:00

Bicubic parametric patches are widely used in various geometric applications. These patches are critical in CAD/CAM systems, which are applied in the automotive industry and mechanical and civil engineering. Commonly, Hermite, Bézier, Coons, or NURBS patches are employed in practice. However, the construction of the Hermite bicubic patch is often not easy to explain formally. This contribution presents a new formal method for constructing the Hermite bicubic plate based on the tensor product approach.

A CNN-RNN hybrid approach for Polish license plate recognition: Harnessing transfer learning and real-world validation

2025-06-23T21:31:29+00:00

Automated license plate recognition (LPR) systems have garnered substantial attention within the field of intelligent transportation systems, owing to their pivotal role in facilitating toll collection, enhancing traffic management, and ensuring operational efficiency. Despite recent breakthroughs in convolutional and recurrent neural network architectures, Polish LPR remains underexplored, with most existing approaches relying on conventional optical character recognition. This study proposes a hybrid convolutional neural network – recurrent neural network (CNN-RNN) model equipped with a Thin-Plate Spline (TPS) transformation module, a ResNet-based feature extractor, a bidirectional Long Short-Term Memory (LSTM) sequence model, and an attention-based decoder to address the unique challenges of Polish license plates. The model is trained on a high-difficulty dataset, comprising real-world images without explicit character-level bounding boxes. Empirical evaluations underscore the efficacy of the proposed system, with competitive accuracy and normalized edit distance scores achieved on Polish, Czech, Hungarian, and Slovak datasets. Additionally, transfer learning from closely related Central European plate formats to Polish data demonstrates marked improvements in convergence and overall performance. Further validation on a challenging video-based dataset reveals the robustness of the proposed approach, evidencing its potential applicability in real-world scenarios and highlighting majority voting as an effective strategy to enhance system reliability under variable conditions.