Object State Identification Using Predicted Image Regions

The patent divides the processing into two separate classifiers: a first classifier that detects object regions in images, and a second classifier that predicts motion trajectories. This segmentation allows each classifier to specialize in a specific task, reducing the overall computational load compared to using a single comprehensive neural network, while maintaining prediction accuracy through coordinated operation of both classifiers.

The first classifier performs preliminary detection of object regions and extracts relevant features before the second classifier performs motion prediction. This preliminary action prepares the data in advance, allowing the second classifier to focus solely on trajectory prediction with reduced computational requirements, thereby decreasing overall processing time while maintaining prediction capability.

The patent segments the comprehensive object analysis into two distinct processing stages handled by separate classifiers. The first classifier handles object detection and region identification, while the second classifier handles motion trajectory prediction. This segmentation reduces the computational complexity required in each individual component, allowing the system to achieve high identification accuracy without requiring excessively complex hardware infrastructure.

The first classifier performs preliminary object detection and extracts features in advance, preparing the data structure and identifying regions of interest before the second classifier performs state identification. This preliminary action enables the second classifier to process only relevant extracted features rather than raw images, significantly improving throughput while keeping the processing logic relatively simple through the structured workflow.