Skip to content
Conference

Efficient Hybrid Models for Multiclass Plant Recognition: From Mask-Guided Focusing to Mobile Deployment Optimization

Aug 2026 · International Conference on Multimedia Analysis and Pattern Recognition · pp. 760-765 · 0 citations · 22 references

Abstract

Plant identification in natural environments is often hindered by complex background clutter, causing accuracy degradation in lightweight models suitable for mobile deployment. We propose a sequential hybrid framework using a mask-guided focusing strategy to isolate specimens and enhance classification accuracy. By integrating segmentation modules such as DeepLabv3, SegFormer-B4, and Mask2Former with classification backbones, our system concentrates on discriminative morphological features while suppressing environmental noise. Evaluation on BigPlants-100, a new curated dataset of 100 Vietnamese plant species, shows that our approach consistently outperforms standalone baselines. Specifically, the combination of ResNet-50 with DeepLabv3 achieved a 92.7% Macro F1-score, marking a 1.0% improvement at 512×512 resolution. Furthermore, hardware-aware optimizations via TensorRT and ONNX Runtime ensure high-speed mobile inference without compromising accuracy. Our findings demonstrate that mask-guided focusing effectively bridges the gap between efficiency and accuracy, providing a robust solution for real-time biodiversity monitoring.

View source

Similar papers

Open access Aug 2026

EnviroSpect-guided ConvMixer–ViT framework for environment-robust potato leaf disease detection

A hybrid deep learning framework that couples ConvMixer, for localized lesionscale feature extraction, with a Vision Transformer, for modeling long-range spatial dependencies, and offers a transferable foundation that could be extended toward scalable, environment-robust plant-disease monitoring in other crops and agri...

A. Sharif, Mudassir Khalil, M. Sajid et al. · 0 citations
Preprint Aug 2026

BoltNet: An Ultra-Lightweight Convolutional Network for On-Device Plant Species Identification

BoltNet is presented, an ultra-lightweight fully convolutional architecture combining a Spatial Redistribution Bottleneck and Logit PreSampling to improve the tradeoff between predictive performance and model size in high-cardinality classification, and the AccuracyCompression Tradeoff is reported as a complementary di...

Daniel Rossi, G. Borghi, R. Vezzani · 0 citations
Open access Sep 2026

Protecting Forests Through Intelligent Edge Computing: CBAM-UNet for Real-Time Deforestation Detection

This study presents a lightweight, attention-enhanced semantic segmentation framework for binary forest canopy segmentation using satellite imagery and provides insights into the trade-offs between accuracy and efficiency in lightweight deep-learning models for remote-sensing applications.

S. P · 0 citations
Open access Aug 2026

Offline Assessment of Stem Detection Technique for Effective Plant Treatment Using Precision Vision Systems

A vision-based approach for detecting plant stems by separating them from surrounding elements such as leaves, soil, and background noise is presented, and the suitability of microprocessor-based platforms for efficient and portable execution of signal processing and machine learning techniques is evaluated.

Mahesh Pawar, Kunal Kulkarni, K. Karandikar et al. · 0 citations
Preprint Sep 2026

PlantC2USeg: Cross-Scale Consistent Pre-Training for Few-Shot Unified Plant Point Cloud Segmentation

Modern crop breeding demands precise organ-level analysis for trait quantification, making plant point cloud segmentation (PPCS) increasingly important. However, conventional deep learning approaches rely heavily on densely annotated datasets that are labor-intensive to acquire. Unified PPCS adaptation from distributio...

Yu Tian, Xintong Jiang, Jan Franklin Adamowski et al. · 0 citations
Open access Aug 2026

DAAL-YOLOv12n: A Method for Detecting Common Forest Pests Incorporating Dynamic Attention and Adaptive Lightweight Modules

Variable illumination, severe occlusion, and the small scale of pest instances in complex forest environments significantly compromise the accuracy of automated forest pest detection. Existing deep learning-based detection models often face challenges in achieving an optimal trade-off among detection accuracy, model li...

Song-Nan Chen, Xue-Lin Chen, Yun Pan · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.