X-DOG: An Intelligent X-ray-based Dangerous Goods Detection and Automatic Alarm System | Semantic Scholar (2024)

Soft-NMS was used to improve the non-maximum suppression of YOLOv4, effectively solving the problem of the high overlap rate of hazardous materials in the X-ray security-inspection dataset and improving accuracy.

The research in this paper focuses on adding additional convolutional layers in the SSD network to re-learn the knowledge of the model learned from the source domain, and shows that compared with the traditional method of fine-tuning, this method has better transfer learning ability on SSD network.

This paper proposes to address the extreme foreground-background class imbalance encountered during training of dense detectors by reshaping the standard cross entropy loss such that it down-weights the loss assigned to well-classified examples, and develops a novel Focal Loss, which focuses training on a sparse set of hard examples and prevents the vast number of easy negatives from overwhelming the detector during training.

OHEM is a simple and intuitive algorithm that eliminates several heuristics and hyperparameters in common use that leads to consistent and significant boosts in detection performance on benchmarks like PASCAL VOC 2007 and 2012.

The proposed network is suitable for sharing the weights in the classifier networks, by which property, the training can be faster with better generalization power, and shows state-of-the-art mAP, which is better than those of the conventional SSD, YOLO, Faster-RCNN and RFCN.

Soft-NMS is proposed, an algorithm which decays the detection scores of all other objects as a continuous function of their overlap with M and improves state-of-the-art in object detection from 39.8% to 40.9% with a single model.

The approach, named SSD, discretizes the output space of bounding boxes into a set of default boxes over different aspect ratios and scales per feature map location, which makes SSD easy to train and straightforward to integrate into systems that require a detection component.

This work introduces a Region Proposal Network (RPN) that shares full-image convolutional features with the detection network, thus enabling nearly cost-free region proposals and further merge RPN and Fast R-CNN into a single network by sharing their convolutionAL features.

An adaptive structural element based on a guided filter kernel is proposed for capturing the local features in infrared images for background suppression and a balanced ring shape is used for two structural elements of top-hat transformation, which can utilize the contrast information between the target and background for target enhancement.

YOLO9000, a state-of-the-art, real-time object detection system that can detect over 9000 object categories, is introduced and a method to jointly train on object detection and classification is proposed, both novel and drawn from prior work.

This paper combines a state-of-the-art classifier with a fast detection framework and augments SSD+Residual-101 with deconvolution layers to introduce additional large-scale context in object detection and improve accuracy, especially for small objects.

...

...