Abstract:This study presents a deep learning research aimed at improving the performance of image classification models and increasing interpretability as well. We explore ways to improve the model by combining attention mechanisms with convolutional neural networks. This study uses garbage classification data in public datasets for training in a supervised learning manner, and employs Grad-CAM (Gradient-weighted Class Activation Mapping) with channel attention mechanism SE (Squeeze-and-Excitation) to generate heat maps for better understanding the decision-making process of the model's classification. By using the Grad-CAM function to generate heat maps, it is possible to visualize the areas which the model focuses on during classification. This provides a method to explain the model's classification decisions, allowing us to better understand the basis of the model's decisions on different categories of images. The model is improved by adding attention modules to different stages of the ResNet50 (Residual Network-50) network, thereby improving the accuracy and performance of the network. We notice that within the same stage, the structure and required attention of each module are consistent, so only one attention module is added in each stage to reduce the burden of network learning and speed up learning. In order to simplify the calculation process of attention, a global tensor is introduced to store the attention of each stage, thereby avoiding repeated calculations in each module. The experimental results show that compared with traditional convolutional neural networks, our proposed method achieves better performance on garbage classification tasks. By combining the attention mechanism and heat map interpretation, our model is able to improve classification accuracy. This is of great significance for image classification tasks in practical applications and helps to promote the research progress of deep learning in interpretability.

Deep Features Analysis with Attention Networks

Deep Neural Networks Evolve Human-like Attention Distribution during Goal-directed Reading Comprehension

Understanding More about Human and Machine Attention in Deep Neural Networks

Deep Learning Methods With the Improved Attention for Explainable Image Recognition

Deep Neural Network with Attention Model for Scene Text Recognition.

HAM: Hybrid Attention Module in Deep Convolutional Neural Networks for Image Classification

Human Vs Machine Attention in Neural Networks: A Comparative Study.

An Empirical Study of Spatial Attention Mechanisms in Deep Networks

Examining Attention Mechanisms in Deep Learning Models for Sentiment Analysis

A review on the attention mechanism of deep learning

What Do Deep Saliency Models Learn about Visual Attention?

Deep Latent Variable Predictive Modeling with Online Bayesian Soft Attention Mechanism

A Comprehensive Review of Deep Neural Network Interpretation Using Topological Data Analysis

Loss-Based Attention for Interpreting Image-Level Prediction of Convolutional Neural Networks.

Where is the Model Looking At?--Concentrate and Explain the Network Attention

Interpreting Attention Models with Human Visual Attention in Machine Reading Comprehension

On the Interpretability of Attention Networks

Attention Flows: Analyzing and Comparing Attention Mechanisms in Language Models

Better Deep Visual Attention with Reinforcement Learning in Action Recognition.

Deep Fashion Analysis with Feature Map Upsampling and Landmark-Driven Attention

Bio-Inspired Representation Learning for Visual Attention Prediction