.png){kind=icon}
.png){kind=icon}
Return
Interpretable morphometric correction-based power model for accurate weight estimation of the sea squirt (Halocynthia roretzi)
.png){kind=icon}
PRE
AI
D
DOI：10.1007/s11694-026-04820-x.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The projected-area (PA)-only model for estimating the body weight of the sea squirt (Halocynthia roretzi) relies solely on a single projected image and fails to adequately account for shape-induced bias. To overcome this limitation, a morphometric correction-based power model (MCPM) was developed in this study by integrating the aspect ratio (AR) and shape factor (SF) into a form-derived correction layer within the PA-based power-function structure. A dataset of 2,433 individuals was analyzed using log-transformed ordinary least squares estimation with Duan smearing correction under a stratified five-fold cross-validation procedure. The results demonstrated that MCPM (Model III) achieved an RMSE of 4.862 g and a MAPE of 6.214%, corresponding to 12.3% and 12.3% reductions, respectively, compared with the conventional PA-only Model I. The positive AR coefficient compensated for the overestimation observed in elongated individuals, whereas the positive SF coefficient compensated for the underestimation found in compact-type individuals, indicating that these morphometric indices functioned not as simple input extensions but as an interpretable, coefficient-based correction mechanism. Because PA, AR, and SF can all be derived from a single-camera projection, MCPM may be implemented as a lightweight correction framework without requiring substantial modification of existing imaging hardware. Accordingly, the proposed approach may support practical application under a single-projection workflow and may be extended through species- and setup-specific recalibration for other morphologically variable aquaculture species.
Keywords:
Morphometric correction-based power model
Projected area
Sea squirt
Shape factor
Weight estimation
Journal
J
IF:
3.3
Papers:
1.3K
Citations:
1.1W
Organization
Cited Papers