Sony

Eight AI and machine learning patents cover player interaction, content creation, and communication systems. Sony's voice control system processes commands from multiple players at once, using microphone arrays to pinpoint who's speaking and route their instructions without requiring anyone to switch controllers. A biometric monitoring patent captures heart rate and other physical signals during gameplay, automatically bookmarking intense moments for spectators and enabling NPCs to adjust their behavior when they detect player stress. Two avatar creation systems take different approaches to the same problem: one transforms user selfies into characters that match a game's visual style, while another combines video-based motion capture with text descriptions to generate custom characters complete with personalized animations. The gesture recognition patent translates sign language and physical movements into chat messages, adjusting the tone and content based on what's happening in the game rather than providing literal translations. An NPC interaction system tracks how players engage with characters across multiple sessions, then uses those patterns to customize future dialogue beyond simple branching choices. The message filtering patent goes further than blocking profanity by rewriting communications to fit each recipient's current gameplay situation while keeping the sender's original meaning intact. A video analysis tool predicts how audiences will respond to game trailers and promotional materials before release, letting studios refine their marketing content based on patterns learned from past campaigns.

The single cloud gaming patent applies different quality levels to objects within the same frame, allocating higher resolution and frame rates to gameplay-critical elements while reducing bandwidth for background details. This approach treats pixels differently based on their importance rather than encoding everything uniformly.

Sony's VR and AR tracking patent addresses the problem of controllers losing visual tracking by automatically switching between camera-based SLAM and motion sensors. The system evaluates which tracking method is working better at any given moment and transitions between them smoothly, preventing the jarring jumps that happen when current headsets suddenly lose sight of a controller.

The Graphics patent modifies character meshes based on what players do repeatedly, making progression visible to others in multiplayer environments. If someone performs certain actions frequently within a time window, their avatar's geometry gradually changes to reflect those behaviors, creating automatic visual feedback that reverses when play patterns shift.

Sony's Audio patent uses AI to separate music vocals from instrumental tracks when dialogue or voice chat starts, adjusting specific frequencies to prevent different voice sources from clashing. The system works in real time without requiring developers to prepare multiple versions of their music assets or manually mix Audio for different scenarios.