Abstract:Recycled and recirculated books, such as ancient texts and reused textbooks, hold significant value in the second-hand goods market, with their worth largely dependent on surface preservation. However, accurately assessing surface defects is challenging due to the wide variations in shape, size, and the often imprecise detection of defects. To address these issues, we propose DDNet, an innovative detection model designed to enhance defect localization and classification. DDNet introduces a surface defect feature extraction module based on a deformable convolution operator (DC) and a densely connected FPN module (DFPN). The DC module dynamically adjusts the convolution grid to better align with object contours, capturing subtle shape variations and improving boundary delineation and prediction accuracy. Meanwhile, DFPN leverages dense skip connections to enhance feature fusion, constructing a hierarchical structure that generates multi-resolution, high-fidelity feature maps, thus effectively detecting defects of various sizes. In addition to the model, we present a comprehensive dataset specifically curated for surface defect detection in recycled and recirculated books. This dataset encompasses a diverse range of defect types, shapes, and sizes, making it ideal for evaluating the robustness and effectiveness of defect detection models. Through extensive evaluations, DDNet achieves precise localization and classification of surface defects, recording a mAP value of 46.7% on our proprietary dataset - an improvement of 14.2% over the baseline model - demonstrating its superior detection capabilities.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper aims to solve the problem of **second - hand book surface defect detection**. Specifically, for recycled and re - circulated books (such as ancient books and second - hand textbooks), the following challenges exist in the accurate assessment of their surface defects: 1. **Diversity in shape, size and position**: The surface defects of these books have various shapes, different sizes, and their positions are difficult to be accurately located. 2. **Limitations of traditional methods**: Traditional manual inspection methods are inefficient and highly subjective, and it is difficult to meet market demands. 3. **Deficiencies of existing deep - learning models**: Existing deep - learning models perform poorly when dealing with complex geometric deformations and multi - scale features. To solve these problems, the author proposes a novel detection model named **DDNet**. The main features of DDNet include: - **Introducing a surface defect feature extraction module based on deformable convolution operator (DC)**: It dynamically adjusts the convolution grid to better capture subtle shape changes and improve boundary division and prediction accuracy. - **Adopting a densely - connected feature pyramid module (DFPN)**: It enhances feature fusion through dense skip connections, constructs multi - resolution, high - fidelity feature maps, and thus effectively detects defects of various sizes. In addition, the author also constructs a dataset specifically for surface defect detection, which contains 6,366 images, covering multiple types of surface defects, such as stains, moth - eaten holes, creases, etc. Experimental results show that the mAP value of DDNet on this dataset reaches 46.7%, which is 14.2% higher than that of the baseline model. ### Summary This paper significantly improves the accuracy and efficiency of second - hand book surface defect detection by proposing the DDNet model and its related technologies, and solves the deficiencies of traditional methods and existing deep - learning models in this field.

DDNet: Deformable Convolution and Dense FPN for Surface Defect Detection in Recycled Books

A Wafer Surface Defect Detection Method Built on Generic Object Detection Network

Automated Detection of Defects with Low Semantic Information in X-ray Images Based on Deep Learning

Joining Spatial Deformable Convolution and a Dense Feature Pyramid for Surface Defect Detection

MDD-Net: A novel defect detection model of material microscope image

A Deep Learning-Based Surface Defect Inspection System Using Multiscale and Channel-Compressed Features

A nondestructive automatic defect detection method with pixelwise segmentation

Coarse-to-Fine Few-Shot Defect Recognition with Dynamic Weighting and Joint Metric

ETDNet: Efficient Transformer-Based Detection Network for Surface Defect Detection

FDSNeT: an Accurate Real-Time Surface Defect Segmentation Network.

Normal Reference Attention and Defective Feature Perception Network for Surface Defect Detection

FDADNet: Detection of Surface Defects in Wood-Based Panels Based on Frequency Domain Transformation and Adaptive Dynamic Downsampling

A Novel Defect Detection System for Complex Freeform Surface Structures

PDD-Net: A Precise Defect Detection Network Based on Point Set Representation

An Efficient Network for Surface Defect Detection

An Adaptive Image Segmentation Network for Surface Defect Detection

Deformation Feature Extraction and Double Attention Feature Pyramid Network for Bearing Surface Defects Detection

ESDDNet: Efficient small defect detection network of workpiece surface

A Pixel-Level Segmentation Convolutional Neural Network Based on Global and Local Feature Fusion for Surface Defect Detection

Decision Fusion Network with Perception Fine-tuning for Defect Classification

DDSNet: Deep Dual-Branch Networks for Surface Defect Segmentation