Small fasteners are widely used in freight train. However, due to the complex working environment, the missing of fasteners always happens, causing traffic accidents and property loss. Thus, detection of the missing of small fasteners is essential for the safety of freight train. This paper takes three critical missing small fasteners as examples: pin-missing, bolt-missing and rivet-missing defects, and proposes a two-stage small defect detection model. First, a copy-pasting method is used to augment the defects in the training image, then the ResNet-101 backbone is adopted to extract defect features, and a feature pyramid network is used to construct feature pyramid. After obtaining defect features, an improved region proposal network is introduced to generate defect proposals. Finally, a fully convolutional neural network is trained to operate pixel-wise segmentation. The experimental results on manually collected fastener defect dataset show that our model possesses high precision and efficiency on the inspection of pin-missing, bolt-missing and rivet-missing defects. Besides, the proposed model could be used for similar small defects detection.

The diagnosis of freight train defects is essential for railway traffic safety. Recently, many freight train defect detection systems based on image processing were developed. However, most of them used hand-engineered features, whose performance was easily degraded by the low image quality and dynamic background. The deep learning method has the merit of extracting robust multi-scale information from images. In this paper, a hierarchical features-based instance detection (HID) model was developed for instance-level defect detections. A deep residual neural network model was firstly used to extract invariant and discriminative features. Then, acquired multi-scale features were processed by a region proposal neural network to generate a coarse defect region and make defect classification. Finally, instance-level predictions were made on the coarse defect region using acquired features. Experiments on the collected dataset that has six types of defects demonstrated the effectiveness of the HID. Additionally, comparisons with the state-of-the-art approaches further demonstrated its promising performance. We also verified its high performance in transfer learning.

Fine-grained fashion retrieval searches for items that share a similar attribute with the query image. Most existing methods utilize a pre-trained feature extractor (e.g. ResNet 50) to capture image representations. However, a pre-trained feature backbone is typically trained for image classification and object detection, which are fundamentally different tasks from fine-grained fashion retrieval. Therefore, existing methods suffer from a feature gap problem when directly using the pre-trained backbone for fine-tuning. To solve this problem, we introduce an attribute-guided multi-level attention network (AG-MLAN). Specifically, we first enhance the pre-trained feature extractor to capture multi-level image embedding, thereby enriching the low-level features within these representations. Then, we propose a classification scheme where images with the same attribute, albeit with different values, are categorized into the same class. This can further alleviate the feature gap problem by perturbing object-centric feature learning. Moreover, we propose an improved attribute-guided attention module for extracting more accurate attribute-specific representations. Our model consistently outperforms existing attention based methods when assessed on the FashionAI (62.8788% in MAP), DeepFashion (8.9804% in MAP), and Zappos50k datasets (93.32% in Prediction accuracy). Especially, ours improves the most typical ASENet_V2 model by 2.12%, 0.31%, and 0.78% points in FashionAI, DeepFashion, and Zappos50k datasets, respectively. The source code is available in https://github.com/Dr-LingXiao/AG-MLAN and will be made public once accepted.