A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is
Although deep convolutional networks have achieved great performance in face recognition tasks, the challenge of domain discrepancy still exists in real world applications. Lack of domain coverage of training data (source domain) makes the learned models degenerate in a testing scenario (target domain). In face recognition tasks, classes in two domains are usually different, so classical domain adaptation approaches, assuming there are shared classes in domains, may not be reasonable solutionsarXiv:2102.13319v1 fatcat:cucw46ezajbs3auondjhogojte
more »... or this problem. In this paper, self-supervised learning is adopted to learn a better embedding space where the subjects in target domain are more distinguishable. The learning goal is maximizing the similarity between the embeddings of each image and its mirror in both domains. The experiments show its competitive results compared with prior works. To know the reason why it can achieve such performance, we further discuss how this approach affects the learning of embeddings.
Using this method, Wu et al. ... diabetes, of Yuanshi Shengmai Chenggu tablets on avascular osteonecrosis of the femoral head  , of Jie-Geng-Tang on lipopolysaccharide-induced acute lung injury in mice  , and of the Tao Hong Si Wu ...doi:10.1186/s13020-018-0221-x pmid:30619503 pmcid:PMC6311004 fatcat:z664ct2tkndthowcekmassense
Wu et al. ...doi:10.3390/s150716981 pmid:26184219 pmcid:PMC4541917 fatcat:ta2olnbwy5anjiekewe2wasaku
It provides facial images with a wide range of ages from newborns to 116, and its images were crawled from search engines such as Google and Bing. ...doi:10.1109/access.2022.3145195 fatcat:zruitvaeazacto2vky3jk3du4y
In this study, a new criterion for determining the number of classes an image should be segmented is proposed. This criterion is based on discriminant analysis for measuring the separability among the segmented classes of pixels. Based on the new discriminant criterion, two algorithms for recursively segmenting the image into determined number of classes are proposed. The proposed methods can automatically and correctly segment objects with various illuminations into separated images fordoi:10.5281/zenodo.1081228 fatcat:enso4eiuvzfablmelfefkv73im
more »... processing. Experiments on the extraction of text strings from complex document images demonstrate the effectiveness of the proposed methods.1
JPEG2000 is a new international standard for still image compression. It provides various functions in one single coding stream and the better compression quality than the traditional JPEG, especially in the high compression ratio. However, the heavy computation and large internal memory requirement still restrict the consumer electronics applications. In this paper, we propose a QCB (quad code block)-based DWT method to achieve the higher parallelism than the traditional DWT approach ofdoi:10.1109/tce.2004.1362517 fatcat:iveejk44g5hznetulqah5xjwpu
more »... 0 coding process. Based on the QCB-based DWT engine, three code blocks can be completely generated after every fixed time slice recursively. Thus, the DWT and EBCOT processors can process simultaneously and the high computational EBCOT then has the higher parallelism of the JPEG2000 encoding system. By changing the output timing of the DWT process and parallelizing with EBCOT, the internal tile memory size can be reduced by a factor of 4. The memory access cycles between the internal tile memory and the code block memory also decrease with the smooth encoding flow.
A general event data recorder is a device installed in automobiles to record information related to vehicle crashes or accidents. The data provide a better understanding of how certain crashes come about. This study made a prototype of a driving behaviour-based event data recorder (DBEDR), which provides the information of driving behaviours and a danger level. The authors approach is to recognise the seven behaviours: normal driving, acceleration, deceleration, changing to the left lane ordoi:10.1049/iet-its.2013.0009 fatcat:oxelhhizczcibocnhmfwnsv3q4
more »... t lane, zigzag driving and approaching the car in front by the hidden Markov models. All data were collected from a real vehicle and evaluated in a real road environment. The experimental results show that the proposed method achieved an average detection ratio of 95% for behaviour recognition. The danger level is inferred by fuzzy rules involved with the above behaviours. DBEDR recorded the recognised driving behaviours and the danger level, and the places were stored with the assistance of a global positioning system receiver. By integrating Google Maps, the locations, the driving behaviour occurrences, the danger level on the travel routes and the recorded images, the proposed DBEDR could be more useful compared with the traditional EDRs.
Then we demonstrated a self-polarizing THz LC phase shifter based on this design. a All correspondences should be addressed to Yan-qing Lu, email@example.com and Biao-bing Jin, firstname.lastname@example.org II. ...doi:10.1063/1.3626560 fatcat:6dsrvx6etvaonn5uj4dptfs2tm
 , Wu et al.  , and Liu et al.  , we utilize evaluations to upweight source images that are visually similar to target images. ...  , Wu et al.  , and Liu et al.  by mining informative images in the training dataset to mitigate domain shift. 1) Domain Discriminator: The architecture from Zhang et al. ...arXiv:2102.13327v2 fatcat:asr5ok6fd5d2jkhvr5wn5j5zma
To contrast the performance of different extensions of the image edges, we incorporate the proposed algorithm into our previous coder based on an enhancing zerotree coding (Wu and Su, 1998). ... That is, circular extensions of an image, we used our previous study, the enchancing zerotree coding (EZC) (Wu and Su, 1998) , to test four images, namely Barbara, Goldhill, Lena, and Mandrill. ...doi:10.1080/02533839.1999.9670455 fatcat:4zuoderl5rh3rij4ucxamjr5tm
This study presents two sets of translation invariant wavelet transforms for coding an arbitrarily shaped image. Each set can be viewed as a shape-adaptive discrete wavelet transform (SA-DWT) with the property of translation invariance. The proposed transform schemes have the following merits: (1) they are translation invariant, (2) no sharp transition appears at the edges of the image, (3) the number of pixels is maintained constant after transformation, (4) the correlation of pixels is fullydoi:10.1016/s0165-1684(99)00104-8 fatcat:etz3mrpeu5gtrfs55rixp5mpsm
more »... xploited, and (5) the property of self-similarity across scales is preserved. 1999 Elsevier Science B.V. All rights reserved. Zusammenfassung Diese Studie praK sentiert zwei translationsinvariante Wavelet-Transformationen fuK r die Kodierung eines beliebig geformten Bildes. Jede davon kann als eine formadaptive diskrete Wavelet-Transformation (SA-DWT) mit der Eigenschaft der Translationsinvarianz angesehen werden. Die vorgeschlagenen Transformationen haben folgende Vorteile: (1) sie sind translationsinvariant, (2) es erscheint kein scharfer U G bergang an den BildraK ndern, (3) die Pixelanzahl bleibt nach der Transformation konstant, (4) die Korrelation der Pixel wird vollstaK ndig ausgenutzt, und (5) die Eigenschaft der SelbstaK hnlichkeit wird fuK r alle Skalen erhalten. 1999 Elsevier Science B.V. All rights reserved. Re2 sume2 Cette eH tude preH sente deux ensembles de transformations en ondelettes invariantes par translation pour le codage d'images de forme arbitraire. Chaque ensemble peut e( tre vu comme une transformation en ondelettes discre'te adaptive en forme (SA-DWT) preH sentant la proprieH teH d'invariance par translation. Les meH thodes de transformation proposeH es ont les avantages suivants: (1) elles sont invariantes par translation, (2) aucune transition abrupte n'apparam( t sur les contours de l'image, (3) le nombre de pixels est maintenu constant apre's transformation, (4) la correH lation des pixels est pleinement exploiteH e, et (5) la proprieH teH d'auto-similariteH au travers des eH chelles est preH serveH e. 1999 Elsevier Science B.V. All rights reserved. 0165-1684/99/$ -see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 -1 6 8 4 ( 9 9 ) 0 0 1 0 4 -8
Journal of Robotics
This paper presents a vision-based technology for localizing targets in 3D environment. It is achieved by the combination of different types of sensors including optical wheel encoders, an electrical compass, and visual observations with a single camera. Based on the robot motion model and image sequences, extended Kalman filter is applied to estimate target locations and the robot pose simultaneously. The proposed localization system is applicable in practice because it is not necessary todoi:10.1155/2011/548042 fatcat:dq2xxg5z3zabfhzhx733haj7li
more »... the initializing setting regarding starting the system from artificial landmarks of known size. The technique is especially suitable for navigation and target tracing for an indoor robot and has a high potential extension to surveillance and monitoring for Unmanned Aerial Vehicles with aerial odometry sensors. The experimental results present "cm" level accuracy of the localization of the targets in indoor environment under a high-speed robot movement.
This paper proposes a new variable step-size sign algorithm (VSSA) for unknown channel estimation or system identification, and applies this algorithm to an environment containing two-component Gaussian mixture observation noise. The step size is adjusted using the gradient-based weighted average of the sign algorithm. The proposed scheme exhibits a fast convergence rate and low misadjustment error, and provides robustness in environments with heavy-tailed impulsive interference.doi:10.1016/j.sigpro.2014.03.030 fatcat:wngyjl55ujagpn4zmomjd6whb4
ObjectivesPrimary aldosteronism (PA) is characterized by the autonomous excessive production of aldosterone in the adrenal cortex. Aldosterone is associated with damages to heart muscle and skeletal muscle. The purpose of this study was to evaluate serum levels of muscle injury markers and their associated factors in patients with primary aldosteronism.MethodsWe retrospectively enrolled subjects with PA and essential hypertension (EH) who had completed testing for serum high sensitivitydoi:10.3389/fendo.2022.799174 pmid:35265036 pmcid:PMC8899019 fatcat:vpod7iewhzapzaujvdq5afow2a
more »... T (hs-TnT), creatine kinase isoenzyme MB (CK-MB) and myoglobin from the database of the Chongqing Primary Aldosteronism Study (CONPASS). Univariate and multivariate linear regression analyses were performed to analyze the influencing factors of myocardial injury markers.ResultsIn total, 278 patients with PA and 445 patients with EH were enrolled in this study. Compared with EH patients, serum concentrations of hs-TnT [7.0 (4.0–12.0) vs. 6.0 (3.0–11.0) ng/L; p=0.005] and myoglobin [24.2 (21.0–38.1) vs. 21.8 (21.0–31.9) μg/L; p=0.023] were significantly higher among PA patients, while no significant difference of CK-MB was found between two groups [1.4 (1.0–2.0) vs. 1.3 (0.9–1.9) μg/L; p=0.154]. Univariate linear regression analysis showed that myoglobin was negatively correlated with serum potassium (β=-0.31; p<0.01) and positively correlated with plasma aldosterone concentration (β=0.40; p<0.01) in the PA group, while no significant correlation was found between hs-TnT and biochemical parameters. After adjusting for multiple confounders, myoglobin was negatively correlated with serum potassium (β=-0.15; p<0.05) and positively correlated with plasma aldosterone concentration (β=0.34; p<0.01) in the PA group.ConclusionsThe serum level of myoglobin was significantly increased in PA patients, and myoglobin was independently correlated with plasma aldosterone concentration.
« Previous Showing results 1 — 15 out of 3,008 results