2nd place in ICDAR 2019 Robust Reading Challenge on Reading Chinese Text on Signboard, only slightly worse than Tencent (0.8150 vs. 0.8144)
2019年, 在美团主办的图像中的中文店铺名识别竞赛第二, 惜败于腾讯 (0.8150 vs. 0.8144)
https://rrc.cvc.uab.es/files/ICDAR2019-ReCTS.pdf
Machine Learning Group
Posted on May 27, 2019 by a1096448
2nd place in ICDAR 2019 Robust Reading Challenge on Reading Chinese Text on Signboard, only slightly worse than Tencent (0.8150 vs. 0.8144)
2019年, 在美团主办的图像中的中文店铺名识别竞赛第二, 惜败于腾讯 (0.8150 vs. 0.8144)
https://rrc.cvc.uab.es/files/ICDAR2019-ReCTS.pdf
Posted on March 14, 2019 by a1096448
https://lp.signate.jp/ai-edge-contest/en/
We (Yifan Liu, Tong He, and Chunhua Shen) attended the AI edge contest organised by The Ministry of Economy, Trade and Industry Japan, and won the 3rd Place.
The task is to create an algorithm to segment the image region corresponding to an object of interest at the pixel level. Images are captured by a camera mounted at the front of a vehicle. 90 teams with 402 participants globally attended the contest.
AI Edge Contest Award
Posted on August 18, 2018 by a1096448
Zifeng Wu, Chunhua Shen, Anton van den Hegnel attended the competition of Nuclei images segmentation task and won the first place!
Leaderboard (showing submitter’s ID):
http://miccai.cloudapp.net/competitions/83#results
Grading and diagnosis of tumors in cancer patients have traditionally been done by examination of tissue specimens under a powerful microscope by expert pathologists. While this process continues to be widely applied in clinical settings, it is not scalable to translational and clinical research studies involving hundreds or thousands of tissue specimens. State-of-the-art digitizing microscopy instruments are capable of capturing high-resolution images of whole slide tissue specimens rapidly. Computer aided segmentation and classification has the potential to improve the tumor diagnosis and grading process as well as to enable quantitative studies of the mechanisms underlying disease onset and progression.
The objective of this challenge is to evaluate and compare segmentation algorithms and to encourage the biomedical imaging community to design and implement more accurate and efficient algorithms. The challenge will evaluate the performance of algorithms for detection and segmentation of nuclei in a tissue image. Participants are asked to detect and segment all the nuclear material in a given set of image tiles extracted from whole slide tissue images.
This challenge uses image tiles from whole slide tissue images to reduce computational and memory requirements. The image tiles are rectangular regions extracted from a set of Glioblastoma and Lower Grade Glioma whole slide tissue images. Nuclei in each image tile in the training set have been manually segmented. Note that the tiles are not of the same size.
Posted on August 18, 2018 by a1096448
Our ResNet38 models are included in Wolfram Neural Net Repo: https://resources.wolframcloud.com/NeuralNetRepository/resources/Ademxapp-Model-A-Trained-on-ImageNet-Competition-Data https://resources.wolframcloud.com/NeuralNetRepository/resources/Ademxapp-Model-A1-Trained-on-ADE20K-Data https://resources.wolframcloud.com/NeuralNetRepository/resources/Ademxapp-Model-A1-Trained-on-Cityscapes-Data https://resources.wolframcloud.com/NeuralNetRepository/resources/Ademxapp-Model-A1-Trained-on-PASCAL-VOC2012-and-MS-COCO-Data
Posted on August 8, 2018 by a1096448
The AIML (Australian Institute for Machine Learning) team, who are Zifeng Wu, Chunhua Shen, Anton van den Hegnel, attended this medical imaging competition and won the 1st place for the task of segmentation. Leaderboard:
Posted on July 22, 2018 by a1096448
We were the 4th place in the JD Fashion Item Search Competition. https://fashion-challenge.github.io/rank.html https://fashion-challenge.github.io/#fashion-item
Posted on May 3, 2018 by a1096448
Adversarial PoseNet: A Structure-aware Convolutional Network for Human Pose Estimation Proc. Int. Conf. Computer Vision 2017
Posted on April 30, 2018 by a1096448
Slides can be accessed at https://cloudstor.aarnet.edu.au/plus/s/Oyqtc4DABsFvddb
Posted on February 19, 2018 by a1096448
This work will be presented at IEEE Conf. Computer Vision and Pattern Recognition (CVPR) 2018.
Posted on February 17, 2018 by a1096448
Detecting individual pedestrians in a crowd remains a challenging problem since the pedestrians often gather together and occlude each other in real-world scenarios. In this paper, we first explore how a state-of-the-art pedestrian detector is harmed by crowd occlusion via experimentation, providing insights into the crowd occlusion problem. Then, we propose a novel bounding box […]
School of Computer Science
Level 4, Ingkarni Wardli
North Terrace Campus
THE UNIVERSITY OF ADELAIDE
ADELAIDE SA 5005 AUSTRALIA