Plant Phenomics / 2022 / Article / Tab 6

Research Article

PSegNet: Simultaneous Semantic and Instance Segmentation for Point Clouds of Plants

Table 6

Quantitative comparison of instance segmentation performances of six networks including PSegNet.

MethodsTobacco leafTomato leafSorghum leafMean

mPrec (%)ASIS78.5480.2179.0479.26
PlantNet87.7485.5079.3984.21
PSegNet (ours)87.8089.9286.6888.13

mRec (%)ASIS56.2764.8472.0864.40
PlantNet69.3676.6381.8375.94
PSegNet (ours)77.0978.6282.1379.28

mCov (%)ASIS62.8876.6174.2671.25
PlantNet71.9883.3482.6379.32
PSegNet (ours)80.8884.8984.2783.35

mWCov (%)ASIS73.9582.7377.3178.00
PlantNet84.8389.4885.6886.66
PSegNet (ours)90.7290.4587.4689.54

The best values are in boldface.