Abstract:Currently, image recognition based on deep neural networks has become the mainstream direction of research; therefore, significant progress has been made in its application in the field of tea detection. Many deep models exhibit high recognition rates in tea leaves detection. However, deploying these models directly on tea-picking equipment in natural environments is impractical; the extremely high parameters and computational complexity of these models make it challenging to perform real-time tea leaves detection. Meanwhile, lightweight models struggle to achieve competitive detection accuracy; therefore, this paper addresses the issue of computational resource constraints in remote mountain areas and proposes Reconstructed Feature and Dual Distillation (RFDD) to enhance the detection capability of lightweight models for tea leaves. In our method, the Reconstructed Feature selectively masks the feature of the student model based on the spatial attention map of the teacher model; it utilizes a generation block to force the student model to generate the teacher's full feature. The Dual Distillation comprises Decoupled Distillation and Global Distillation. Decoupled Distillation divides the reconstructed feature into foreground and background features based on the Ground-Truth. This compels the student model to allocate different attention to foreground and background, focusing on their critical pixels and channels. However, Decoupled Distillation leads to the loss of relation knowledge between foreground and background pixels. Therefore, we further perform Global Distillation to extract this lost knowledge. Since RFDD only requires loss calculation on feature map, it can be easily applied to various detectors. We conducted experiments on detectors with different frameworks, using a tea dataset collected at the Huangshan Houkui Tea Plantation. The experimental results indicate that, under the guidance of RFDD, the student detectors have achieved performance improvements to varying degrees. For instance, a one-stage detector like RetinaNet (ResNet-50) experienced a 3.14% increase in Average Precision (AP) after RFDD guidance. Similarly, a two-stage model like Faster RCNN (ResNet-50) obtained a 3.53% improvement in AP. This offers promising prospects for lightweight models to efficiently perform real-time tea leaves detection tasks.

Enhanced Knowledge Distillation for Advanced Recognition of Chinese Herbal Medicine

An Optimized Deep Convolutional Neural Network for Dendrobium Classification Based on Electronic Nose

Image recognition of traditional Chinese medicine based on deep learning

Knowledge Distillation Method for Surface Defect Detection.

Image Recognition of Chinese herbal pieces Based on Multi-task Learning Model

Rapid Mold Detection in Chinese Herbal Medicine Using Enhanced Deep Learning Technology

Image Recognition of Chinese Herbal Medicine Using Adaptive Gamma Correction Based on Convolutional Neural Network

Multiple attentional pyramid networks for Chinese herbal recognition

Knowledge Distillation Facilitates the Lightweight and Efficient Plant Diseases Detection Model

Research on Image Classification of Chinese Medicinal Materials Powder Based on Texture Extraction and Improved Lightweight Network

RSKD: Enhanced medical image segmentation via multi-layer, rank-sensitive knowledge distillation in Vision Transformer models

Learning lightweight tea detector with reconstructed feature and dual distillation

Chinese Herbal Recognition based on Competitive Attentional Fusion of Multi-hierarchies Pyramid Features

Combined Channel Attention and Spatial Attention Module Network for Chinese Herbal Slices Automated Recognition

CCNNet: a novel lightweight convolutional neural network and its application in traditional Chinese medicine recognition

EDKSANet: An Efficient Dual-Kernel Split Attention Neural Network for the Classification of Tibetan Medicinal Materials

Target Detection Model Distillation Using Feature Transition and Label Registration for Remote Sensing Imagery

ClearKD: Clear Knowledge Distillation for Medical Image Classification

A novel Chinese herbal medicine classification approach based on EfficientNet

MSKD: Structured knowledge distillation for efficient medical image segmentation

An instance-based deep transfer learning method for quality identification of Longjing tea from multiple geographical origins