Vision Target Tracker Based on Incremental Dictionary Learning and Global and Local Classification

Algorithm 1

Algorithm of Fisher discrimination dictionary learning.

Input: 𝐼_𝑡 is the tracking result of prior frame, {𝐼¹_𝑡+1,𝐼²_𝑡+1,…,𝐼^𝑛_𝑡+1} is set of

candidate samples credible positions of the center-point coordinates in next

frame. 𝐷 is Over-complete dictionary, 𝑇 is frame numbers.

(1) for 𝑡=1∶𝑇

(2) SRGC

  calculate 𝑎_GC = argmin_𝑎GC𝑦−𝐷𝛼_GC²₂+𝛾𝛼_GC1

(3) calculate identity(𝑦)=argmin_𝑖{𝑒_𝑖}; where 𝑦={𝐼¹_𝑡+1,𝐼²_𝑡+1,…,𝐼^𝑛_𝑡+1},

(4) obtaining the center point (𝑥,𝑦) from 𝐼^𝑖_(𝑡+1)

(5) obtaining the 𝐼¹_{(𝑡+1)𝑥,𝑦},𝐼²_{(𝑡+1)𝑥,𝑦},…,𝐼^𝑛_{()𝑡+1𝑥,𝑦} by (𝑥,𝑦) and Affine transformation of 𝐼_𝑡

(6) SRLC to 𝐼¹_{(𝑡+1)𝑥,𝑦},𝐼²_{(𝑡+1)𝑥,𝑦},…,𝐼^𝑛_{()𝑡+1𝑥,𝑦};

  calculate 𝑎_LC=argmin_𝑎LC{𝑦−𝐷𝛼_LC²₂+𝛾₁𝛼_LC1+𝛾₂𝛼_GC−𝑋²₂};

(7) calculate identity(𝑦)=argmin_𝑖{𝑒_𝑖}, where 𝑒=𝑦−𝐷𝑎_LC²₂+𝛾₁𝛼_LC1

Output: 𝐼^𝑖_{(𝑡+1)(𝑥,𝑦)}