Crosstalk Cascades - GitHub Pages · Crosstalk Cascades for Frame-Rate Pedestrian Detection...
1
Crosstalk Cascades for Frame-Rate Pedestrian Detection Wolf Kienzle Microso( Research Redmond [email protected] Piotr Dollár Microso( Research Redmond [email protected] Ron Appel California Ins4tute of Technology [email protected] Crosstalk versus so( cascades (sweep over γ) Classify using dense sliding window (4px steps) Classifier is Boosted depth-2 trees Neighboring windows are correlated (average classifier responses around true posi4ves) Baseline So( Cascades • FAST: 3060 fps (530x speedup) • ACCURATE: stateoftheart detec4on • ROBUST: consistent performance across datasets • GENERAL: applicable to any mul4stage detector • TRAINABLE: no extra supervised data required • CODE: feature computa4on code available at: hOp://vision.ucsd.edu/~pdollar/toolbox/doc/index.html see paper for more details Less jitter - or - Smaller K More jitter - or - Larger K Tradeoff between Performance and Neighborhood Correla4on • Use: jiOer = ±2, K = 4096, Neighborhood of [7x7x3] (12px steps) Summary Don’t Ignore thy Neighbors Previous State-of-the-Art Overview PROBLEM: detec4on is computa4onally demanding OBSERVATION: adjacent windows are evaluated independently, no informa4on is shared IDEA: exploit neighborhood correla4ons during cascade: combine excita(on and inhibi(on GAINS: 530 x speedup over standard cascades SPEED: 3060 fps on 640x480px images (1 core, no GPU) Integral Channel Features using FPDW for fast scale pyramid Crosstalk Cascades Effec4veness of soI cascades (sweep over γ) Parameter (γ) Controls SpeedAccuracy Tradeoff Performance on other datasets hOp://www.vision.caltech.edu/Image_Datasets/CaltechPedestrians/ Additional Pedestrian Datasets Caltech Pedestrian Dataset ETH Pedestrian Dataset TUDBrussels Dataset Crosstalk Cascades Accuracy and speedup of crosstalk cascades Speedup versus classifier complexity Excitatory Cascades Inhibitory Cascades Speedup versus classifier complexity Speedup versus classifier complexity A(er k stages in the Boos4ng classifier, neighbors are excited if: H k >θ E k A(er k stages in the Boos4ng classifier, an evalua4ng window is inhibited if it has a neighboring window N such that: H k /H N k <θ I k Accuracy and speedup of so( cascades with constant threshold θ * and varying thresholds θ R k Reduces Computation for All Windows Combina4on of SoI, Excitatory, and Inhibitory cascades, reducing computa4on across the board Stateoftheart detec-on accuracy at framerate speeds! Wow! Logaverage miss rate (MR) versus speed for various detectors on INRIA pedestrians Fast and Accurate Results false posi4ves per image false posi4ves per image
Transcript of Crosstalk Cascades - GitHub Pages · Crosstalk Cascades for Frame-Rate Pedestrian Detection...
Crosstalk Cascades for Frame-Rate Pedestrian Detection
Wolf Kienzle Microso( Research Redmond
Piotr Dollár Microso( Research Redmond
Ron Appel California Ins4tute of Technology
Crosstalk versus so( cascades (sweep over γ)
Classify using dense sliding window (4px steps) Classifier is Boosted depth-2 trees
Neighboring windows are correlated (average classifier responses around true posi4ves)
Baseline So( Cascades
• FAST: 30-‐60 fps (5-‐30x speedup) • ACCURATE: state-‐of-‐the-‐art detec4on • ROBUST: consistent performance across datasets • GENERAL: applicable to any mul4-‐stage detector • TRAINABLE: no extra supervised data required • CODE: feature computa4on code available at: hOp://vision.ucsd.edu/~pdollar/toolbox/doc/index.html
see paper for more details
Less jitter - or -
Smaller K
More jitter - or -
Larger K
Tradeoff between Performance and Neighborhood Correla4on
• Use: jiOer = ±2, K = 4096, Neighborhood of [7x7x3] (12px steps)
Summary
Don’t Ignore thy Neighbors
Previous State-of-the-Art
Overview PROBLEM: detec4on is computa4onally demanding OBSERVATION: adjacent windows are evaluated independently, no informa4on is shared IDEA: exploit neighborhood correla4ons during cascade: combine excita(on and inhibi(on GAINS: 5-‐30 x speedup over standard cascades SPEED: 30-‐60 fps on 640x480px images (1 core, no GPU)
Integral Channel Features using FPDW for fast scale pyramid
Crosstalk Cascades
Effec4veness of soI cascades (sweep over γ)
Parameter (γ) Controls Speed-‐Accuracy Tradeoff
Performance on other datasets hOp://www.vision.caltech.edu/Image_Datasets/CaltechPedestrians/
Additional Pedestrian Datasets
Caltech Pedestrian Dataset ETH Pedestrian Dataset TUD-‐Brussels Dataset
Crosstalk Cascades
Accuracy and speedup of crosstalk cascades
Speedup versus classifier complexity
Excitatory Cascades Inhibitory Cascades
Speedup versus classifier complexity Speedup versus classifier complexity
A(er k stages in the Boos4ng classifier, neighbors are excited if: Hk > θEk
A(er k stages in the Boos4ng classifier, an evalua4ng window is inhibited if it has a neighboring window N such that: Hk/HN
k < θIk
Accuracy and speedup of so( cascades with constant threshold θ* and varying thresholds θRk
Reduces Computation for All Windows
Combina4on of SoI, Excitatory, and Inhibitory cascades, reducing computa4on across the board
State-‐of-‐the-‐art detec-on accuracy
at frame-‐rate speeds! Wow!
Log-‐average miss rate (MR) versus speed for various detectors on INRIA pedestrians
Fast and Accurate Results
false posi4ves per image false posi4ves per image
