AI Science Tracker
EN

Tomato seedling stem and leaf segmentation method based on an improved ResNet architecture

College of Information Technology, Jilin Agricultural University, Changchun, China; College of Information Engineering, Changchun University of Finance and Economics, Changchun, China · 28 Aug 2025 · biology

Discovery

Tomato seedling stem and leaf segmentation method based on an improved ResNet architecture

Source check: Source checked. Not independent scientific confirmation.

Finding: The proposed X-ResNet segmented tomato seedling stems and leaves from 3D point clouds and supported phenotypic measurement. The study reports 95.11% training accuracy, 3.26 percentage points above ResNet18, and manual-measurement correlations of R² 0.941 for plant height, 0.752 for stem diameter, 0.945 for leaf area and 0.943 for leaf inclination; reported absolute percentage errors were below 6%. The abstract does not provide a held-out segmentation accuracy.

Technology: X-ResNet point-cloud segmentation with bottleneck, downsampling and specialized convolutional layers

Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.

AI contribution
The learned X-ResNet segmented tomato stems and leaves from point clouds; the resulting height, diameter, area and inclination estimates were compared with manual measurements.
Test and result
X-ResNet point-cloud segmentation and extracted plant traits were compared with manual measurements. The paper reports trait-specific R² values and errors, 95.11% training accuracy, and an inference time of 4.02 seconds versus 5.37 seconds for ResNet18.
Limitation / still unknown
The 95.11% value is training accuracy; the abstract does not state held-out segmentation accuracy. Transfer to other cultivars or growing conditions is not established.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC12444385/

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.

Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..

No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.

Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Open in the full catalog for filters, charts and related discoveries. Browse biology discoveries.