[GitHub] HarDNet for ImageNet (Pytorch)
35% faster than ResNet

[GitHub] SSD-HarDNet for Object Detection (Pytorch)
SSD512-HarDNet85 achieves mAP 35.1 in COCO test-dev, which is even faster than SSD512-VGG16

[GitHub] FC-HarDNet for Semantic Segmentation (Pytorch)

FC-HarDNet-70 achieves mIoU 76 in Cityscapes dataset with 70fps (Titan V), standing the first place of real-time semantic segmentation algorithms shown in Papers with Code

Hardnet Result



US Patents


1. Chao; Ping et al., “Merged access units in frame buffer compression,US Patent 10,418,002 (2017)

Merged access units in frame buffer compression patent


2. Chuang; Tzu-Der and Chao; Ping et al., “Method and apparatus for resource sharing between intra block copy mode and inter prediction mode in video coding systems,” US Patent 10,397,588 (2017)

3. Chao; Ping et al., “Method and apparatus of loop filters for efficient hardware implementation,” US Patent 10,306,246 (2016)

4. Liu; Tsu-Ming and Chao; Ping et al.,“Method and apparatus for accessing compressed data and/or uncompressed data of image frame in frame buffer,” US Patent 9,767,529 (2014)

5. Chao; Ping et al., “Method and apparatus for video processing incorporating deblocking and sample adaptive offset,” US Patent 9,635,360 (2013)

6. Lin; Huei-Min and Chao; Ping et al.,“Video processing system with shared/configurable in-loop filter data buffer architecture and related video processing method thereof”, US Patent 9,438,911 (2013)

7. Chao; Ping et al., “Method and apparatus for sample adaptive offset in a video decoder”, US Patent 9,344,717 (2013)



US Patents Under Exam


1. Chao; Ping et al., “Method and Device for Storing an Image in to a Memory”, US Patent Application 20180107616 (2017)

2. CHIU; Min-Hao and Chao; Ping et al., “Decoding System for Tile-based Videos”, US Patent Application 20180139464 (2017)

3. Lin; Hsiu-Yi and Chao; Ping et al., “Method and Apparatus of Bandwidth Estimation and Reduction for Video Coding,” US Patent Application 20170188033 (2016)

4. Chao; Ping et al.,Simulated Transparent Device,US Patent Application 20170032559 (2016)

5. Chao; Ping et al.,Method and System with Data Reuse in Inter-Frame Level Parallel Decoding,” US Patent Application 20160191935 (2015)

6. Chao; Ping et al.,Mixed-Level Multi-core Parallel Video Decoding System,” US Patent Application 20160191922 (2015)



1. Ping Chao, Chao-Yang Kao, Yu-Shan Ruan, Chien-Hsiang Huang, and Youn-Long Lin.  “HarDNet: A Low Memory Traffic Network,” in International Conference on Computer Vision (ICCV), 2019.

2. Ju et al. (Mediatek Group Work), “A 2.6mm2 0.19nJ/pixel VP9 and multi-standard decoder LSI for Android 4K TV applications,” 42nd European Solid-State Circuits Conference (ESSCIRC) 2016, Pages: 109 – 112

3. Ju et al. (Mediatek Group Work), “A 0.5 nJ/Pixel 4 K H.265/HEVC Codec LSI for Multi-Format Smartphone Applications,” IEEE Journal of Solid-State Circuits (JSSC) 2016, Volume: 51, Issue: 1, Pages: 56 – 67

4. Ju et al. (Mediatek Group Work), “Energy and area efficient hardware implementation of 4K Main-10 HEVC decoder in Ultra-HD Blu-ray player and TV systems,” IEEE International Conference on Multimedia and Expo (ICME) 2015, Pages: 1 – 6

5. Ju et al. (Mediatek Group Work), “18.6 A 0.5nJ/pixel 4K H.265/HEVC codec LSI for multi-format smartphone applications,” IEEE International Solid-State Circuits Conference – (ISSCC) Digest of Technical Papers 2015, Pages: 1 – 3

6. Ju et al. (Mediatek Group Work), “A 0.2nJ/pixel 4K 60fps Main-10 HEVC decoder with multi-format capabilities for UHD-TV applications,” 40th European Solid State Circuits Conference (ESSCIRC) 2014, Pages: 195 – 198

7. Ju et al. (Mediatek Group Work), “A 4K×2K@60fps multi-standard TV SoC processor with integrated HDMI/MHL receiver,” Symposium on VLSI Circuits Digest of Technical Papers 2014, Pages: 1 – 2

8. Ju et al. (Mediatek Group Work), “A 775-µW/fps/view H.264/MVC decoder chip compliant with 3D Blu-ray specifications,” IEEE International Symposium on Circuits and Systems (ISCAS) 2012, Pages: 1440 – 1443

9. Ping Chao; Youn-Long Lin, “An elastic software cache with fast prefetching for Motion Compensation in video decoding,” IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis (CODES+ISSS) 2010, Pages: 23 – 32

10. Ping Chao; Youn-Long Lin,“Reference frame access optimization for ultra high resolution H.264/AVC decoding,” IEEE International Conference on Multimedia and Expo (ICME) 2008, Pages: 1441 – 1444

11. Ping Chao; Youn-Long Lin,“A motion compensation system with a high efficiency reference frame pre-fetch scheme for QFHD H.264/AVC decoding,” IEEE International Symposium on Circuits and Systems (ISCAS) 2008, Pages: 256 – 259