Search

Your search keyword '"Wu, Junyu"' showing total 316 results
Start Over Author "Wu, Junyu"
316 results on '"Wu, Junyu"'

32. Automatically Building Face Datasets of New Domains from Weakly Labeled Data with Pretrained Models

Author

Ding, Shengyong, Wu, Junyu, Xu, Wei, and Chao, Hongyang

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Training data are critical in face recognition systems. However, labeling a large scale face data for a particular domain is very tedious. In this paper, we propose a method to automatically and incrementally construct datasets from massive weakly labeled data of the target domain which are readily available on the Internet under the help of a pretrained face model. More specifically, given a large scale weakly labeled dataset in which each face image is associated with a label, i.e. the name of an identity, we create a graph for each identity with edges linking matched faces verified by the existing model under a tight threshold. Then we use the maximal subgraph as the cleaned data for that identity. With the cleaned dataset, we update the existing face model and use the new model to filter the original dataset to get a larger cleaned dataset. We collect a large weakly labeled dataset containing 530,560 Asian face images of 7,962 identities from the Internet, which will be published for the study of face recognition. By running the filtering process, we obtain a cleaned datasets (99.7+% purity) of size 223,767 (recall 70.9%). On our testing dataset of Asian faces, the model trained by the cleaned dataset achieves recognition rate 93.1%, which obviously outperforms the model trained by the public dataset CASIA whose recognition rate is 85.9%.

Published
2016

33. Deep Joint Face Hallucination and Recognition

Author

Wu, Junyu, Ding, Shengyong, Xu, Wei, and Chao, Hongyang

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Deep models have achieved impressive performance for face hallucination tasks. However, we observe that directly feeding the hallucinated facial images into recog- nition models can even degrade the recognition performance despite the much better visualization quality. In this paper, we address this problem by jointly learning a deep model for two tasks, i.e. face hallucination and recognition. In particular, we design an end-to-end deep convolution network with hallucination sub-network cascaded by recognition sub-network. The recognition sub- network are responsible for producing discriminative feature representations using the hallucinated images as inputs generated by hallucination sub-network. During training, we feed LR facial images into the network and optimize the parameters by minimizing two loss items, i.e. 1) face hallucination loss measured by the pixel wise difference between the ground truth HR images and network-generated images; and 2) verification loss which is measured by the classification error and intra-class distance. We extensively evaluate our method on LFW and YTF datasets. The experimental results show that our method can achieve recognition accuracy 97.95% on 4x down-sampled LFW testing set, outperforming the accuracy 96.35% of conventional face recognition model. And on the more challenging YTF dataset, we achieve recognition accuracy 90.65%, a margin over the recognition accuracy 89.45% obtained by conventional face recognition model on the 4x down-sampled version., Comment: 10 pages, 2 figures

Published
2016

43. Enhancing Geological Understanding and Identifying Gold Anomalies in the Ailaoshan Orogen.

Author

ZHAO, Shiyu, YANG, Lin, SONG, Yiwei, DONG, Yuntao, FENG, Lihao, LI, Huajian, WU, Junyu, and WANG, Qingfei

Subjects
WATERSHEDS, RIVER sediments, GOLD, STANDARD deviations, FACTOR analysis, OROGENIC belts, MORPHOTECTONICS
Abstract

The Ailaoshan Orogen in the southeastern Tibet Plateau, situated between the Yangtze and Simao blocks, underwent a complex structural, magmatic, and metamorphic evolution resulting in different tectonic subzones with varying structural lineaments and elemental concentrations. These elements can conceal or reduce anomalies due to the mutual effect between different anomaly areas. Dividing the whole zone into subzones based on tectonic settings, ore cluster areas, or sample catchment basins (Scb), geochemical and structural anomalies associated with gold (Au) mineralization have been identified utilizing mean plus twice standard deviations (Mean + 2STD), factor analysis (FA), concentration‐area (C‐A) modeling of stream sediment geochemical data, and lineament density in both the Ailaoshan Orogen and the individual subzones. The FA in the divided 98 Scbs with 6 Scbs containing Au deposits can roughly ascertain unknown rock types, identify specific element associations of known rocks and discern the porphyry or skarn‐type Au mineralization. Compared with methods of Mean + 2STD and C‐A model of data in the whole orogen, which mistake the anomalies as background or act the background as anomalies, the combined methods of FA and C‐A in the separate subzones or Scbs works well in regional metallogenic potential analysis. Mapping of lineament densities with a 10‐km circle diameter is not suitable to locate Au deposits because of the delineated large areas of medium‐high lineament density. In contrast, the use of circle diameters of 1.3 km or 1.7 km in the ore cluster scale delineates areas with a higher concentration of lineament density, consistent with the locations of known Au deposits. By analyzing the map of faults and Au anomalies, two potential prospecting targets, Scbs 1 and 63 with a sandstone as a potential host rock for Au, have been identified in the Ailaoshan Orogen. The use of combined methods in the divided subzones proved to be more effective in improving geological understanding and identifying mineralization anomalies associated with Au, rather than analyzing the entire large area. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results