Point Cloud Decoding with Wiener Filtering for Reconstruction Quality

The patent applies preliminary action by performing filtering on the reconstructed point cloud during the encoding process. The encoder calculates a filtering coefficient based on the initial point cloud and reconstructed point cloud, then applies this filter to reduce distortions before encoding. This preliminary filtering action improves prediction accuracy for subsequent blocks while adding controlled complexity to the encoding process.

The patent implements feedback by using the reconstructed point cloud to calculate the filtering coefficient, which is then fed back into the encoding process. The filtering flag information and filtering coefficient are signaled in the bitstream, allowing the decoder to apply the same filter. This feedback mechanism ensures that the filtering improves prediction accuracy for future predictions while managing encoding complexity through selective application.

The patent applies parameter changes by dynamically adjusting the filtering process based on calculated filtering coefficients. The filtering coefficient is determined based on the initial point cloud and reconstructed point cloud characteristics, allowing the system to adapt the filtering strength and type to the specific content. This enables high reconstruction quality while minimizing processing time by applying strong filtering only when needed.

The patent implements partial action by selectively applying filtering based on the filtering flag information. The system can choose to apply filtering to certain blocks or regions where it is most beneficial, rather than uniformly applying it to the entire point cloud. This partial application strategy maintains reconstruction quality for critical areas while reducing overall processing time.

The patent applies the taking out principle by extracting only the essential filtering information (filtering flag and filtering coefficient) from the complex filtering process and encoding only these key parameters in the bitstream. This allows the decoder to reconstruct the filtered point cloud without transmitting the entire filtering process, reducing bitstream complexity while maintaining distortion reduction effectiveness.

The patent uses copying by having the encoder calculate and signal the filtering coefficient, which is then copied and applied at the decoder. This ensures that both encoder and decoder use the same filtering parameters, maintaining consistency and reliability in distortion reduction while avoiding the need to transmit complex filtering algorithms, thus managing bitstream complexity.