high-level-feature

UniWorld: High-Resolution Semantic Encoders for Unified Visual Understanding and Generation

This work proposes a feature refined end-to-end tracking framework with a balanced performance using a high-level feature refine tracking framework. The feature refine module enhances the target feature representation power that allows the network to capture salient information to locate the target. The attention module is employed inside the feature refine mechanism to improve network discrimination power that augments the network ability to track the target in challenging scenarios.

HierbaNetV1 is a novel convolutional neural network (CNN) architecture with a cutting-edge feature extraction technique

ML classifier trained with raw pixels as features and high level features and then comparing their accuracies