670 likes | 675 Views
A Vision-Based Driver Assistance System Based on Dynamic Visual Model. 指導教授:陳世旺博士 傅楸善博士. 研 究 生:方瓊瑤. Outline. Introduction Dynamic visual model (DVM) Neural modules Road sign recognition system System to detect changes in driving environments
E N D
A Vision-Based Driver Assistance System Based on Dynamic Visual Model 指導教授:陳世旺博士 傅楸善博士 研 究 生:方瓊瑤
Outline • Introduction • Dynamic visual model (DVM) • Neural modules • Road sign recognition system • System to detect changes in driving environments • System to detect motion of nearby moving vehicles • Conclusions
Introduction (1) -- ITS • Intelligent transportation system (ITS) • Advanced traffic management systems (ATMS) • Advanced traveler information systems (ATIS) • Commercial vehicle operations (CVO) • Advanced public transportation systems (APTS) • Advanced rural transportation systems (ARTS) • Advanced vehicle control and safety systems (AVCSS) • Driver assistance systems (DAS)
Introduction (2) -- DAS • Driver assistance systems (DAS) • Safety • Passive • Active • Driving is a sophisticated process • The technology of vehicle • The temperament of the driver
Introduction (3) -- VDAS • Vision-based driver assistance systems (VDAS) • Difficulties of VDAS • Weather and illumination • Daytime and nighttime • Vehicle motion and camera vibration
Introduction (4) • Subsystems of VDAS • Road sign recognition system • System to detect changes in driving environments • System to detect motion of nearby moving vehicles
Introduction (5) -- DVM • DVM: dynamic visual model • A computational model for visual analysis using video sequence as input data • Two ways to develop a visual model • Biological principles • Engineering principles • Artificial neural networks
Video images Data transduction Sensory component Episodic Memory Information acquisition Spatialtemporal information Perceptual component STA neural module No Focuses of attention Yes Feature detection Categorical features Conceptual component CART neural module Category Pattern extraction Patterns CHAM neural module Action Dynamic Visual Model
Physical stimuli Data compression Transducer Low-level feature extraction Sensory analyzer High-level feature extraction Perceptual analyzer Classification and recognition Conceptual analyzer Class of input stimuli Human Visual Process
Neural Modules • Spatial-temporal attention (STA) neural module • Configurable adaptive resonance theory (CART) neural module • Configurable heteroassociative memory (CHAM) neural module
STA Neural Network (1) ak ai Output layer (Attention layer) nk ni Inhibitory connection wij Excitatory connection xj nj Input layer
Gaussian function G Attention layer ni rk nk corresponding neurons wkj nj Input neuron The linking strengths between the input and the attention layers STA Neural Network (2) • The input to attention neuron nidue to input stimuli x:
Interaction + Lateral distance “Mexican-hat” function of lateral interaction STA Neural Network (3) • The input to attention neuron ni due to lateral interaction:
STA Neural Network (4) • The net input to attention neuron ni : : a threshold to limit the effects of noise where 1< d <0
STA Neural Network (5) stimulus activation t 1 1 p pd The activation of an attention neuron in response to a stimulus.
Orienting subsystem Attentional subsystem Category representation field F2 y Signal generator Reset signal S Input representation field F1 + q + + r + - + - - + + G p + G + + G G - v + + + u + + x - + G + w + Input vector i ART2 Neural Network (1) CART
ART2 Neural Network (2) • The activities on each of the six sublayers on F 1: where I is an input pattern where where the J th node on F 2 is the winner
ART2 Neural Network (3) • Initial weights: • Top-down weights: • Bottom-up weights: • Parameters:
v1 v2 vi vn Output layer (Competitive layer) i Excitatory connection wij xj j Input layer HAM Neural Network (1) CHAM
HAM Neural Network (2) • The input to neuron nidue to input stimuli x: nc: the winner after the competition
Objective • Get information about road • Warn drivers • Enhance traffic safety • Support other subsystems
Conceptual Component— Classification results of CART Training Set Test Set
Conceptual Component— Training Patterns for CHAM
Discussion • Vehicle and camcorder vibration • Incorrect recognitions Input patterns Recognition results Correct patterns
Definition • The environmental changes in expressways: • Left-lane-change • Right-lane-change • Tunnel-entry • Tunnel-exit • Expressway-entry • Expressway-exit • Overpass-ahead
Objective • Coordinate DAS subsystems • Update parameters • Detect unexpected changes • Detect rapid changes
Discussion • Curved roads • Shadows • Multiple environmental changes
Introduction • Motions of the Vehicles • Lane change • Speed change • Objective • Simple motion detection • Complex motion detection
Improved DVM • Two problems: • The motions of vehicles may occur anywhere on the road. • Training a CART neural network to recognize various complex motions is quite difficult. • Solutions: • Feature extraction • Attention map partition • Collection of classification results • Temporal integral process
Attention maps Windowing b1 b2 bn-1 bn Feature extraction Feature extraction Feature extraction Feature extraction CART1 CART2 CARTn-1 CARTn st1 st2 stn-1 stn Decision making No Confirm? Yes Output Flowchart for Conceptual Component
b5 b4 b1 b2 b3 Attention Map Partition
1 2 3 4 5 6 7 8 9 10 gi1 i i gi1 1 2 3 4 5 6 7 8 9 10 Feature Extraction (1) ---Skewness features