待研究论文

短命女 2022-02-17 03:07 297阅读 0赞

1、CenterNet（无anchor 检测：bbox keypoint） 已完成  
知乎：[https://zhuanlan.zhihu.com/p/62836519][https_zhuanlan.zhihu.com_p_62836519]  
论文：[https://arxiv.org/pdf/1904.07850.pdf][https_arxiv.org_pdf_1904.07850.pdf]  
github：[https://github.com/xingyizhou/CenterNet][https_github.com_xingyizhou_CenterNet]  
同时还有corner net：  
博文：[https://mp.weixin.qq.com/s/DoN-vha1H-2lHhbFOaVS8w][https_mp.weixin.qq.com_s_DoN-vha1H-2lHhbFOaVS8w]  
paper path：[https://arxiv.org/pdf/1904.03797v1.pdf][https_arxiv.org_pdf_1904.03797v1.pdf]

fcos、retina、focal loss

2、morphNet（加速网络）  
博文：[https://mbd.baidu.com/newspage/data/landingsuper?context=\{"nid"%3A"news\_9720362154603662894"\}&n\_type=0&p\_from=1][https_mbd.baidu.com_newspage_data_landingsuper_context_nid_3A_news_9720362154603662894_n_type_0_p_from_1]  
github：[https://github.com/google-research/morph-net][https_github.com_google-research_morph-net]

3、octConv（新卷积方式）  
博文：[https://mbd.baidu.com/newspage/data/landingsuper?context=\{"nid"%3A"news\_9720362154603662894"\}&n\_type=0&p\_from=1][https_mbd.baidu.com_newspage_data_landingsuper_context_nid_3A_news_9720362154603662894_n_type_0_p_from_1]  
论文：[https://export.arxiv.org/pdf/1904.05049][https_export.arxiv.org_pdf_1904.05049]  
github：[https://github.com/terrychenism/OctaveConv][https_github.com_terrychenism_OctaveConv]

4、YOLACT: Real-time Instance Segmentation（realtime分割）  
博文：[https://blog.csdn.net/jiaoyangwm/article/details/89176767][https_blog.csdn.net_jiaoyangwm_article_details_89176767]  
github：[https://github.com/dbolya/yolact][https_github.com_dbolya_yolact]

5、TimeCycle（无监督跟踪和光流新方法）  
博文：[https://mp.weixin.qq.com/s/g0tgdRXtR-i1Hf6v6-r1zw][https_mp.weixin.qq.com_s_g0tgdRXtR-i1Hf6v6-r1zw]  
论文：Learning Correspondence from the Cycle-Consistency of Time  
github：[https://github.com/xiaolonw/TimeCycle][https_github.com_xiaolonw_TimeCycle]  
Slides (google drive): [drive.google.com/file/d][drive.google.com_file_d]  
Slides (百度网盘): [https://pan.baidu.com/share/init?surl=prNthUokiqRPELs8J4O-vQ][https_pan.baidu.com_share_init_surl_prNthUokiqRPELs8J4O-vQ]

6、2DASL（3D人脸重建与密集对齐，比PRNet更优秀）  
博文：[https://mp.weixin.qq.com/s/h8tk28IEuqB7g1e-vsyN5g][https_mp.weixin.qq.com_s_h8tk28IEuqB7g1e-vsyN5g]  
论文：Joint 3D Face Reconstruction and Dense Face Alignment from A Single Image with 2D-Assisted Self-Supervised Learning  
[https://arxiv.org/pdf/1903.09359.pdf][https_arxiv.org_pdf_1903.09359.pdf]  
github：[https://github.com/XgTu/2DASL][https_github.com_XgTu_2DASL]

7、高级语义特征提取（检测精度更高，主要针对行人检测）  
论文：High-level Semantic Feature Detection:A New Perspective for Pedestrian Detection  
paper path：[https://arxiv.org/abs/1904.02948v1][https_arxiv.org_abs_1904.02948v1]  
github：[https://github.com/liuwei16/CSP][https_github.com_liuwei16_CSP]

8、实时-视频流检测：  
博文：[https://mp.weixin.qq.com/s/Sq3dBuU8aY3Ug9NBZMc5lA][https_mp.weixin.qq.com_s_Sq3dBuU8aY3Ug9NBZMc5lA]  
论文：[https://arxiv.org/abs/1903.10172][https_arxiv.org_abs_1903.10172]  
github：[https://github.com/tensorflow/models/tree/master/research/lstm\_object\_detection][https_github.com_tensorflow_models_tree_master_research_lstm_object_detection]

9、ThunderNet（旷视、更快检测）  
博文：[https://mp.weixin.qq.com/s/F0ykoKv027ycinsAZZjbWQ][https_mp.weixin.qq.com_s_F0ykoKv027ycinsAZZjbWQ]  
论文：ThunderNet: Towards Real-time Generic Object Detection

efficent net

10、ZQCNN vs libfacedetection（快速人脸检测）  
博文：[https://mp.weixin.qq.com/s/DkvQ86tgU4DarsmY\_qTLdg][https_mp.weixin.qq.com_s_DkvQ86tgU4DarsmY_qTLdg]  
github：[https://github.com/zuoqing1988/ZQCNN-MTCNN-vs-libfacedetection][https_github.com_zuoqing1988_ZQCNN-MTCNN-vs-libfacedetection]  
github2：[https://github.com/ShiqiYu/libfacedetection][https_github.com_ShiqiYu_libfacedetection]

人脸对比可以使用faiss库进行微秒级别的搜索

人脸识别：argface、mtcnn

[https_zhuanlan.zhihu.com_p_62836519]: https://zhuanlan.zhihu.com/p/62836519
[https_arxiv.org_pdf_1904.07850.pdf]: https://arxiv.org/pdf/1904.07850.pdf
[https_github.com_xingyizhou_CenterNet]: https://github.com/xingyizhou/CenterNet
[https_mp.weixin.qq.com_s_DoN-vha1H-2lHhbFOaVS8w]: https://mp.weixin.qq.com/s/DoN-vha1H-2lHhbFOaVS8w
[https_arxiv.org_pdf_1904.03797v1.pdf]: https://arxiv.org/pdf/1904.03797v1.pdf
[https_mbd.baidu.com_newspage_data_landingsuper_context_nid_3A_news_9720362154603662894_n_type_0_p_from_1]: https://mbd.baidu.com/newspage/data/landingsuper?context=%7B%22nid%22%3A%22news_9720362154603662894%22%7D&n_type=0&p_from=1
[https_github.com_google-research_morph-net]: https://github.com/google-research/morph-net
[https_export.arxiv.org_pdf_1904.05049]: https://export.arxiv.org/pdf/1904.05049
[https_github.com_terrychenism_OctaveConv]: https://github.com/terrychenism/OctaveConv
[https_blog.csdn.net_jiaoyangwm_article_details_89176767]: https://blog.csdn.net/jiaoyangwm/article/details/89176767
[https_github.com_dbolya_yolact]: https://github.com/dbolya/yolact
[https_mp.weixin.qq.com_s_g0tgdRXtR-i1Hf6v6-r1zw]: https://mp.weixin.qq.com/s/g0tgdRXtR-i1Hf6v6-r1zw
[https_github.com_xiaolonw_TimeCycle]: https://github.com/xiaolonw/TimeCycle
[drive.google.com_file_d]: http://drive.google.com/file/d
[https_pan.baidu.com_share_init_surl_prNthUokiqRPELs8J4O-vQ]: https://pan.baidu.com/share/init?surl=prNthUokiqRPELs8J4O-vQ
[https_mp.weixin.qq.com_s_h8tk28IEuqB7g1e-vsyN5g]: https://mp.weixin.qq.com/s/h8tk28IEuqB7g1e-vsyN5g
[https_arxiv.org_pdf_1903.09359.pdf]: https://arxiv.org/pdf/1903.09359.pdf
[https_github.com_XgTu_2DASL]: https://github.com/XgTu/2DASL
[https_arxiv.org_abs_1904.02948v1]: https://arxiv.org/abs/1904.02948v1
[https_github.com_liuwei16_CSP]: https://github.com/liuwei16/CSP
[https_mp.weixin.qq.com_s_Sq3dBuU8aY3Ug9NBZMc5lA]: https://mp.weixin.qq.com/s/Sq3dBuU8aY3Ug9NBZMc5lA
[https_arxiv.org_abs_1903.10172]: https://arxiv.org/abs/1903.10172
[https_github.com_tensorflow_models_tree_master_research_lstm_object_detection]: https://github.com/tensorflow/models/tree/master/research/lstm_object_detection
[https_mp.weixin.qq.com_s_F0ykoKv027ycinsAZZjbWQ]: https://mp.weixin.qq.com/s/F0ykoKv027ycinsAZZjbWQ
[https_mp.weixin.qq.com_s_DkvQ86tgU4DarsmY_qTLdg]: https://mp.weixin.qq.com/s/DkvQ86tgU4DarsmY_qTLdg
[https_github.com_zuoqing1988_ZQCNN-MTCNN-vs-libfacedetection]: https://github.com/zuoqing1988/ZQCNN-MTCNN-vs-libfacedetection
[https_github.com_ShiqiYu_libfacedetection]: https://github.com/ShiqiYu/libfacedetection