Kechen Song(Associate professor)


  • Supervisor of Doctorate Candidates  Supervisor of Master's Candidates
  • Name (English):Kechen Song
  • E-Mail:
  • Education Level:With Certificate of Graduation for Doctorate Study
  • Gender:Male
  • Degree:博士
  • Status:Employed
  • Alma Mater:东北大学
  • Teacher College:机械工程与自动化学院


Open time:..

The Last Update Time:..

Switch language:CN

Mobile Version
Current position: Home > DATABASE

Few-shot Surface Defect

    Few-shot Surface Defect Datasets

    1.Classification (Cls):


    1.1 Classification-Strip Steel: FSC-20 Dataset

    We constructed a large-scale few-shot classification dataset named FSC-20(few-shot classification contains 20 types of strip steel surface defects). This dataset was collected from the three datasets the NEU-CLS, the X-SDD, and the GC10-DET, with a total of 20 categories. Specifically, it includes 10 types of cold-rolled defects (all from GC10-DET) and 10 types of hot-rolled defects (6 types from NEU-CLS and the other 4 types from X-SDD).

    Wenli Zhao,  Kechen Song,  Yanyan Wang, Shubo Liang, Yunhui Yan. FaNet: Feature-aware Network for Few Shot Classification of Strip Steel Surface Defects [J].  Measurement,  2023, 208, 112446. (IF:5.6)  (paper)  (Dataset and code

    1.2 Classification-Metal: MSD-Cls Dataset

    We constructed a few-shot metal surface defect dataset named MSD-Cls. The MSD-Cls dataset includes ten types of steel surface defects and ten types of aluminum surface defects. The ten types of steel surface defects are from Northeastern University classification (NEU-CL) datasets and railway surface defect datasets (RSDDs). Each type of steel defect contains 50 images of size 224 × 224. The ten types of aluminum surface defects come from the dataset provided by the 2018 Tianchi competition, the size of these images is 256×256, and the number of pictures in each type ranges from 22 to 48.

    Weiwei Xiao, Kechen Song, Jie Liu and Yunhui Yan.Graph Embedding and Optimal Transport for Few-Shot Classification of Metal Surface Defect [J]. IEEE Transactions on Instrumentation and Measurement2022, 71,5010310.(IF:5.6)   (paper) (Dataset and code


    2.Semantic Segmentation:


    2.1 Semantic Segmentation-Strip Steel: FSSD-12 Dataset

    We constructed a novel few-shot segmentation dataset, FSSD-12, to address the severely insufficient pixel-wise labeled S3D samples in existing works. There are twelve S3D classes in FSSD-12, including abrasionmask, iron-sheet ash, liquid, oxide-scale, oil-spot, water-spot, patch, punching, red-iron sheet, roll-printing, scratch, and inclusion. All raw defective images are integrated from DET GC-10, X-SDD, SD-900, and Surface Defects-4i , which are taken in the production and manufacturing stages. Subsequently, we meticulously annotate the overall defective images with pixel-wise labels.

    Hu Feng, Kechen Song,  Wenqi Cui, Yiming Zhang, Yunhui Yan. Cross Position Aggregation Network for Few-shot Strip Steel Surface Defect Segmentation [J]. IEEE Transactions on Instrumentation & Measurement2023, 72, 5007410. (IF:5.6)    (paper) (Dataset and code

    2.2 Semantic Segmentation-Metal: Surface Defects-4i Dataset

    We constructed a novel dataset, Surface Defects-4i , that uses images and annotations of multiple metal surface defect dataset. In addition, we also add two classes of nonmetal as an extension and further prove the generality of our method. Specifically, before experimenting, we grayscale all images and uniform size to 200 × 200 for ensuring consistency. There are a total of 12 different classes of surface defects in the dataset. Each class includes defective images, groundtruth (GT), and a large number of normal images.

    Yanqi Bao,  Kechen Song, Jie Liu, Yanyan Wang, Yunhui Yan, Han Yu, Xingjie Li. Triplet- Graph Reasoning Network for Few-shot Metal Generic Surface Defect Segmentation [J].IEEE Transactions on Instrumentation and Measurement2021, 70, 5011111(IF:5.6)    (paper (Dataset and code)

Variable Illumination RGB-T Dataset

    A  Variable Illumination RGB-T Dataset: VI-RGBT1500

    Some example images.jpg

    (a)Images under sufficient illumination. (b) Images under uneven illumination. (c) Images under insufficient illumination.


    A Variable Illumination Dataset: VI-RGBT1500

    We propose a variable illumination dataset named VI-RGBT1500 for RGBT image SOD. This is the first time that different illuminations are taken into account in the construction of the RGBT SOD dataset. Three illumination conditions, which are sufficient illumination, uneven illumination and insufficient illumination, are adopted to collect 1500 pairs of RGBT images.

     Kechen Song, Liming Huang, et al.  Multiple Graph Affinity Interactive Network and A Variable Illumination Dataset for RGBT Image Salient Object Detection [J]. IEEE Transactions on Circuits and Systems for Video Technology, 2023.    (paper)(code & dataset)

Visible-Depth-Thermal Image Dataset

    A Novel Visible-Depth-Thermal Image Dataset of Salient Object Detection

    1.jpg 2.jpg

    Visual perception plays an important role in industrial information field, especially in robotic grasping application. In order to detect the object to be grasped quickly and accurately, salient object detection (SOD) is employed to the above task. Although the existing SOD methods have achieved impressive performance, they still have some limitations in the complex interference environment of practical application. To better deal with the complex interference environment, a novel triple-modal images fusion strategy is proposed to implement SOD for robotic visual perception, namely visible-depth-thermal (VDT) SOD. Meanwhile, we build an image acquisition system under variable lighting scene and construct a novel benchmark dataset for VDT SOD (VDT-2048 dataset). Multiple modal images will be introduced to assist each other to highlight the salient regions. But, inevitably, interference will also be introduced. In order to achieve effective cross-modal feature fusion while suppressing information interference, a hierarchical weighted suppress interference (HWSI) method is proposed. The comprehensive experimental results prove that our method achieves better performance than the state-of-the-art methods.  
     Kechen Song, et al.  A Novel Visible-Depth-Thermal Image Dataset of Salient Object Detection for Robotic Visual Perception [J]. IEEE/ASME Transactions on Mechatronics, 2022.  (paper) (Dataset & Code

NEU surface defect database

    NEU surface defect database

    Kechen Song and Yunhui Yan


    In the Northeastern University (NEU) surface defect database, six kinds of typical surface defects of the hot-rolled steel strip are collected, i.e., rolled-in scale (RS), patches (Pa), crazing (Cr), pitted surface (PS), inclusion (In) and scratches (Sc). The database includes 1,800 grayscale images: 300 samples each of six different kinds of typical surface defects.

    Fig. 1 shows the sample images of six kinds of typical surface defects, the original resolution of each image is 200×200 pixels. From Fig. 1, we can clearly observe that the intra-class defects existing large differences in appearance, for instance, the scratches (the last column) may be horizontal scratch, vertical scratch, and slanting scratch, etc. Meanwhile the inter-class defects have similar aspects, e.g., rolled-in scale, crazing, and pitted surface. In addition, due to the influence of the illumination and material changes, the grayscale of the intra-class defect images is varied. In short, the NEU surface defect database includes two difficult challenges, i.e., the intra-class defects existing large differences in appearance while the inter-class defects have similar aspects, the defect images suffer from the influence of illumination and material changes.

    NEU surface defect database.jpg

                                                                                                       Fig.1 NEU-CLS

    For defect detection task, we provided annotations which indicate the class and location of a defect in each image. We have carefully clicked annotations of each target in these images. Fig. 2 shows some examples of detection results on NEU-DET. For each defect, the yellow box is the bounding box indicating its location and the green label is the class score.

    NEU surface defect database-Fig.2.png

    Fig.2 NEU-DET

    How to use the database:   
        If you are using the NEU surface defect database for defect classification task, you only need to download the image-only file, called NEU-CLS.
        If you are using the NEU surface defect database for defect detection task, you need to download the image-with-annotations file, called NEU-DET.

         For Original NEU surface defect databaseGoogle.pngGoogle-Drive, OrBaidu.png Baidu-Pan
         For NEU-CLSGoogle.pngGoogle-Drive, Or Baidu.pngBaidu-Pan
         For NEU-CLS-64:Google.pngGoogle-Drive, Or Baidu.pngBaidu-Pan
         For NEU-DET:Google.pngGoogle-Drive, Or Baidu.pngBaidu-Pan(code:pmqx)

        For   SD-saliency-900 database


       We would appreciate it if you cite our works when using the database:

          Yanqi Bao, Kechen Song, Jie Liu, Yanyan Wang, Yunhui Yan, Han Yu, Xingjie Li, “Triplet- Graph Reasoning Network for Few-shot Metal Generic Surface Defect Segmentation,” IEEE Transactions on Instrumentation and Measuremente,  2021.(paper)
        K. Song and Y. Yan, “A noise robust method based on completed local binary patterns for hot-rolled steel strip surface defects,” Applied Surface Science, vol. 285, pp. 858-864, Nov. 2013.

       Yu He, Kechen Song, Qinggang Meng, Yunhui Yan, “An End-to-end Steel Surface Defect Detection Approach via Fusing Multiple Hierarchical Features,” IEEE Transactions on Instrumentation and Measuremente,  2020,69(4),1493-1504.(paper)



        Vision-based Inspection System for Steel Surface Defects
        For Micro surface defect database:Google.pngGoogle-Drive, OrBaidu.pngBaidu-Pan
        For Oil pollution defect database:Google.pngGoogle-Drive, Or Baidu.pngBaidu-Pan

Micro surface defect database

    Micro surface defect database:

              Micro surface defect database:Google.pngGoogle-Drive, OrBaidu.pngBaidu-Pan


Oil pollution defect database

    Oil pollution defect database:

              Oil pollution defect database:Google.pngGoogle-Drive, Or Baidu.pngBaidu-Pan 

    Oil pollution defect database.jpg