基于Dice系数的弱选择回溯匹配追踪算法

Weak-Selection Backtracking Matching Pursuit Algorithm Based on Dice Coefficient

为进一步提高压缩感知重构算法的重构成功率和重构精度,从原子匹配准则和预选阶段原子选择方式的角度出发,提出一种基于Dice系数的弱选择回溯匹配追踪(weak-selection backtracking matching pursuit based on Dice coefficient,DWBMP)算法.首先,采用Dice系数匹配准则度量两个向量之间的相似性,选出最匹配的原子,以优化支撑集;然后,结合回溯思想和弱选择思想剔除相似性较小的原子,完成预选阶段原子的二次筛选.MATLAB仿真结果显示,相同条件下,DWBMP算法较经典的压缩感知重构算法具有更优的重构精度和重构成功率.

In order to further improve the success rate and accuracy of reconstruction of the compressed sensing reconstruction algorithm,a weak-selection backtracking matching pursuit based on Dice coefficient(DWBMP)algorithm is proposed from the perspective of atomic matching criteria and pre-selection stage’s atom selection methods.First,the Dice coefficient matching criterion is used to measure the similarity between two vectors,and the best matching atom is selected to optimize the support set.Then,the backtracking idea is combined with weak-selection idea to eliminate the atoms with small similarity,thus completing the secondary selection of the atoms in the pre-selection stage.The MATLAB simulation results show that under the same conditions,the DWBMP algorithm has better success rate and accuracy of reconstruction than the classic compressed sensing reconstruction algorithm.