Camouflage Detection An introduction Presented by Ani Starrenburg
Camouflage Detection An introduction Presented by: Ani Starrenburg
General Camouflaging Strategies n Cryptic Camouflage Little Button Quail Traditional US Army Camouflage Pattern
General Camouflaging Strategies n Mimicry Dronefly Rose Greenbow, Confederate Spy
General Camouflaging Strategies n Disruption Sumatran Tiger Dazzle Camouflage
General Camouflaging Strategies n Countershading Impala Non-Countershaded Warship
General Camouflaging Strategies n Translucence/Transparency Seawasp Invisibility Cloak
Detecting Camouflaged Objects:
Camouflage Detection Methods n Standard Object Detection Methods n Edge Detection Models n Contrast Energy Detection Model n Motion Detection n Correlation Models n Gradient Models n Energy Models
Edge Detectors: Gaussian Gradient Laplacian With Gaussian
Canny Detector n Optimal Edge Detector n Multiple Stage Algorithm Perform Gaussian smoothing n Find edge strengths |G| = |Gx| + |Gy| n Detection of edge direction theta = invtan(Gy/Gx) n Relate edge direction to a direction that can be traced in an image n Apply non-maximum suppression n Use hysteresis to eliminate streaking n
La. Placian or Lo. G n Smooth with a Gaussian mask n Calculate the second derivatives n Search for zero crossings Or n Convolve the image with the Laplacian of the Gaussian
Contrast Energy (CE) Model n Uses the output signal from similarly-oriented odd o[x] and even e[x] filters. n Energy function is defined as: E 2(x) = e 2(x) + o 2(x) n Always positive n Shows high output when o(x), e(x) or both are high.
Camouflage Detection Methods to be Discussed n Convexity-Based Detection – exploits the principle of countershading to detect camouflaged objects n Texture Detection – intensive texture analysis distinguishes camouflaged object from background. Also, uses Canny detector to bring up edges
Motion Breaks Camouflage Region of common velocity is perceived As a unit and stands out against the static background
Reichardt Correlation Model n Computes motion as the ratio of the partial derivatives of the input image brightness with respect to space and time. n Two spatially-separate detectors. n Output of one of the detectors is delayed. n The two outputs are multiplied to determine if there is a correlation.
Multichannel Gradient Model n Uses multiple channels of higher derivatives n The more derivatives used lowers the chance of that all will be zero at the same time n Uses a least sqaures approximation of the derivatives
Motion Energy Model n Uses two sets of oriented detectors(leftwards and rightwards), each composed of an odd an even filter. n Energy is calculated by summing the squares of the two similarly-oriented filters. n Calculate opponent energy (difference of leftward and rightward results) n Normalize by dividing by static energy to give velocity estimates
An aside: Research on Active Camouflage n Animals that can escape edge detection n Animals that can camouflage motion
To Do List: n Apply edge detectors and contrast energy detectors to camouflaged and illusory images and view results. n Research visual models developed from observing animal behavior and development. n Research studies in psychology for further understanding of vision process.
Is there a core visual system? C A M O U F L A G E A R T
Bibliography n n n Motion Illusions and Active Camouflage, Lewis Dartnell , http: //www. ucl. ac. uk/~ucbplrd/motion_middle. html Canny Edge Detection Tutorial, Bill Green, http: //www. pages. drexel. edu/~weg 22/can_tut. html Honeybee, http: //www. gpnc. org/honeybee. htm Ground-dwelling birds, http: //www. birdobservers. org. au/ground_birds. htm Sumatran tiger, http: //www. saczoo. com/3_kids/20_camouflage/camouflage_disruptive. htm Biomimicry, http: //www. wordspy. com/words/biomimicry. asp Countershading, http: //www. shipcamouflage. com/ships 2_3_43_countershading. htm Translucence, http: //www. gla. ac. uk/ibls/DEEB/teg/project_pages/counter_shading. htm Canny Edge Detection, http: //homepages. inf. ed. ac. uk/rbf/CVonline/LOCAL_COPIES/OWENS/LECT 6/node 2. html Optical Camouflage, http: //projects. star. t. u-tokyo. ac. jp/projects/MEDIA/xv/VRIC 2003. pdf Multi-Channel Gradient Model, http: //www. psychol. ucl. ac. uk/pmco/Mc. GM. html
- Slides: 21