By Frédéric Cao, José-Luis Lisani, Jean-Michel Morel, Pablo Musé, Frédéric Sur

Recent years have visible dramatic development suit reputation algorithms utilized to ever-growing snapshot databases. they've been utilized to snapshot sewing, stereo imaginative and prescient, photograph mosaics, sturdy item reputation and video or internet photograph retrieval. extra essentially, the facility of people and animals to discover and realize shapes is likely one of the enigmas of notion.

The publication describes a whole strategy that begins from a question snapshot and a picture database and yields an inventory of the photographs within the database containing shapes found in the question photograph. A fake alarm quantity is linked to every detection. Many experiments will exhibit that conventional basic shapes or pictures can reliably be pointed out with fake alarm numbers starting from 10-5 to under 10-300.

Technically talking, there are major concerns. the 1st is extracting invariant form descriptors from electronic photos. the second one is finding out no matter if form descriptors are identifiable because the comparable form or now not. A perceptual precept, the Helmholtz precept, is the cornerstone of this selection.

These judgements depend on basic stochastic geometry and compute a fake alarm quantity. The reduce this quantity, the safer the identity. the outline of the techniques, the numerous experiments on electronic photos and the straightforward proofs of mathematical correctness are interlaced with a view to make a interpreting available to numerous audiences, akin to scholars, engineers, and researchers.

Show description

Read Online or Download A Theory of Shape Identification PDF

Similar stochastic modeling books

Stochastic Dominance: Investment Decision Making under Uncertainty (Studies in Risk and Uncertainty)

This publication is dedicated to funding decision-making below uncertainty. The publication covers 3 easy techniques to this strategy: the stochastic dominance method; the mean-variance technique; and the non-expected application process, targeting prospect thought and its changed model, cumulative prospect concept.

Swarm Intelligence (The Morgan Kaufmann Series in Evolutionary Computation)

Conventional equipment for developing clever computational platforms haveprivileged inner most "internal" cognitive and computational tactics. Incontrast, Swarm Intelligence argues that humanintelligence derives from the interactions of people in a social worldand additional, that this version of intelligence may be successfully utilized toartificially clever structures.

Simulation, Fifth Edition

The fifth version of Ross's Simulation maintains to introduce aspiring and training actuaries, engineers, laptop scientists and others to the sensible features of making automated simulation reports to research and interpret genuine phenomena. Readers learn how to observe result of those analyses to difficulties in a large choice of fields to procure powerful, exact options and make predictions approximately destiny results.

Council for African American Researchers in the Mathematical Sciences: Volume III

This quantity offers study and expository papers awarded on the 3rd and 5th conferences of the Council for African American Researchers within the Mathematical Sciences (CAARMS). The CAARMS is a gaggle devoted to organizing an annual convention that showcases the present study basically, yet now not completely, of African americans within the mathematical sciences, together with arithmetic, operations study, information, and desktop technological know-how.

Extra resources for A Theory of Shape Identification

Example text

The choice of µ could depend on the application. Detected edges may be used for different purposes, for instance shape recognition or image matching. If |Du| is less than 1, then the position of level lines may be locally inaccurate. Eliminating pieces of curves with a gradient smaller than µ = 1 for all images is therefore not restrictive in shape recognition applications. 5 shows an example of the clean up procedure. 26 2 Extracting Meaningful Curves from Images Fig. 5 Meaningful boundary clean up.

Middle: Evian, 481 level lines. 2 Experiments 47 standard PC. When images do not show long level lines, the computation time is less than a second. Fig. 4 Flat parts detection: Bretagne. 1004 detections. Flat parts as small as the ones in the letters of the name of the street are detected (about 10 pixels high). Flat parts in the boundaries of the shadows can be eliminated by dropping the probability threshold, as can be seen on Fig. 5. 11 show the result of the proposed flat parts detector over all level lines in an image.

One can therefore figure out that at least two parameters are involved in a flatness measurement. One measures the length of the flat part and the other gives the amplitude of the oscillations. Thus, the flatness definition problem can be viewed as the question F. , A Theory of Shape Identification. Lecture Notes in Mathematics 1948. c Springer-Verlag Berlin Heidelberg 2008 41 42 3 Robust Shape Directions Fig. 1 A piece of discrete curve with the underlying chord C (thick segment line) of reducing two parameters to a more abstract one, the flatness.

Download PDF sample

Rated 4.63 of 5 – based on 31 votes