Target detection, industrial defects, image segmentation -- deep learning data set induction

90+ Deep learning, open source data set sorting ｜ Including target detection 、 Industrial defects 、 Image segmentation and other directions . Tweet gives 95 Deep learning open source datasets . It includes 11 A direction （ Small target detection 、 object detection 、 Industrial testing 、 Face recognition 、 Human posture estimation 、 Autopilot 、 Target tracking 、 Image classification 、 Image recognition 、 Image segmentation 、NLP relevant ） And the corresponding open source links . The convenience of data sets is provided in the research process of our problems . meanwhile , The diverse datasets show interesting and diverse directions for deep learning tasks .

1. Small target detection （AI-TOD Aerial image data set 、iSAID Large scale data set of aerial images 、TinyPerson Data sets 、Deepscores Data sets 、 Dense pedestrian detection data set 、 Caltech pedestrian detection data set 、NWPU VHR-10 Satellite image data set 、Inria Aerial image data set 、RSOD Remote sensing image data sets 、 Small target detection data set ）

2. Target detection open source data set （COCO2017 Data sets 、 Flame and smoke image data set 、DOTA Aerial image data set 、AITEX Data sets 、T-LESS Data sets 、H²O Pedestrian interaction detection data set 、SpotGarbage Garbage identification data set 、NAO Nature confronts sample data sets 、Labelme Image data set 、 Indian vehicle data set 、Seeing 3D chairs Chair detection model 、SUN09 Scenario understanding dataset 、Unsplash Image retrieval dataset 、HICO-DET Character interaction detection data set 、 Shanghai University of science and technology demographic data set 、 Domestic waste data set 、RMFD Masks mask the face dataset 、GTSRB German traffic sign data set 、VOC2005 Vehicle data set 、Winegrape Test data set 、 Global wheat detection data set 、Linkopings Traffic sign data set 、 Protective equipment - Helmet and vest inspection 、 Caltech camera trap dataset 、 Underwater waste detection data set ）

3. Industrial test data set （ Data set of pothole detection 、 Tianchi aluminum profile surface defect data set 、Kylberg Texture dataset 、 Steel strip surface defect data set of Northeast University 、Severstal Strip defect data set 、UCI Strip defect data set 、DAGM 2007 Data sets 、RSDDs Rail surface defect data set 、KTH-TIPS Texture image dataset 、 pcb （PCB） Defect dataset ）

4. Open source data sets are commonly used in face recognition （IMDB-WIKI Face data set 、WiderFace Face detection dataset 、LFW Portrait image data set 、GENKI Face image dataset 、 Columbia University public figures facial database 、CelebA Face data set 、 US Department of defense face database 、MTFL Face recognition dataset 、BioID Face data set 、PersonID Face recognition dataset 、CMU PIE Face database 、Youtube Video face dataset 、CASIA Face image dataset 、Caltech Face database ）

5. Human posture estimation （MPII Manikin dataset 、MPII Human posture data set 、KTH Multi view football dataset 、 Pennsylvania action dataset 、BBC Attitude data set 、Poser Data sets 、 field 3D Pose dataset 、V-COCO Data sets 、 ikea ASM Data sets 、 Stereo human pose estimation data set 、AIST++ Dance movement dataset 、HiEve Data sets ）

6. Autopilot （KITTI Road dataset 、CrackForest Data sets 、KITTI-2015 Stereo dataset 、KITTI-2015 Optical flow data set 、.KITTI-2015 Scene stream dataset 、KITTI Depth data sets ）

7. Target tracking （ALOV300++ Trace data set 、UCF50 Action recognition data set 、SBU Kinect Interactive dataset ）

8. Image classification data set （ PET image data set 、 Cat dataset 、 Stanford dog dataset 、CBCL Street scene data 、Stanford Car picture data 、MNIST Handwritten digital image data set 、Kaggle Garbage classification image data set ）

9. Image recognition data set （ Street view house number (SVHN) Data sets 、3D MNIST Digital recognition image data ）

10. Image segmentation data set （LVIS Data sets 、 High density crowd and moving object video data set 、DAVIS Video segmentation data set ）

11.NLP Related data sets （ Document photocopying and content data ）

Specific links of data sets can be queried ：https://mp.weixin.qq.com/s/N2bFlRTTNeXTgJ1rMsL5Ag