Search
Search Funnelback University
- Refined by:
- Date: 2022
1 -
42 of
42
search results for TALK:PC53 20 |u:mi.eng.cam.ac.uk
where 0
match all words and 42
match some words.
Results that match 1 of 2 words
-
Kate Knill - Biography
mi.eng.cam.ac.uk/~kmk/bio.html18 Mar 2022: As Languages Manager (2000 - 2002), she led a cross-site team that developed over 20 languages for speech recognition and speaker verification. -
wxRegSurf
mi.eng.cam.ac.uk/~ahg/wxRegSurf/sslm.html12 Sep 2022: mc3_6756_lmarked.ply individual.ply n/a n/a 0 0 0 999 0 20 10 1 -1 0 0 1 0 # review canonical_mc3_6756_lmarked.ply individual.ply -
wxRegSurf
mi.eng.cam.ac.uk/~ahg/wxRegSurf/femur.html12 Sep 2022: To perform the nonrigid registration, we set the Iterations slider to around 20 and select Registration->SimilarityLAD (preferable) or Registration->SimilaritySSM (if in a rush), before pressing Start. ... the lesser trochanters are approximately level, -
wxRegSurf
mi.eng.cam.ac.uk/~ahg/wxRegSurf/lrhr.html12 Sep 2022: hr_femur_thickness.bin 0 0 20 999 0 0 1 0 -1 1 0 0 0. ... reg. file. The final line of the above example loads the two sets of cortical data, performs 20 smoothing cycles on the HR data, applies the similarity transformation from the. -
wxRegSurf
mi.eng.cam.ac.uk/~ahg/wxRegSurf/vertebra.html12 Sep 2022: To perform the nonrigid registration, we set the Iterations slider to around 20, check the Auto sequence checkbox and select Registration->SimilarityLAD (preferable) or Registration->SimilaritySSM (if in a rush), before ... Next, check the Auto sequence -
Scaling digital screen reading with one-shot learning and…
mi.eng.cam.ac.uk/~cipolla/archive/Publications/inproceedings/2021-WACV-scaling-digital-meters.pdf9 Apr 2022: 0. 20. 40. 60. 80. 100. Prec. isio. n (%. ). Multimeter. 0 5 10 15 20 25Norm dist from GT (px). ... 20] E. Rublee, V. Rabaud, K. Konolige, and G. Bradski. Orb: Anefficient alternative to sift or surf. -
R. MECCA ET. AL : LUCES: NEAR-FIELD PHOTOMETRIC STEREO ...
mi.eng.cam.ac.uk/~cipolla/publications/inproceedings/2021-BMVC-LUCES-photometric-stereo-dataset.pdf9 Apr 2022: Xiong et al. [41] have proposed a dataset of 7 objects using 20 directionallights calibrated with two chrome spheres. ... 4 ExperimentsIn this section, we evaluate four competing near-field methods namely [18, 20, 32, 35]. -
Estimating and Exploiting the Aleatoric Uncertaintyin Surface Normal…
mi.eng.cam.ac.uk/~cipolla/publications/inproceedings/2021-ICCV-surface-normal-uncertainty.pdf9 Apr 2022: We assumethat the uncertainty is heteroscedastic [20] (i.e. certain pix-els have higher uncertainty than the others). ... 5. [20] Alex Kendall and Yarin Gal. What uncertainties do weneed in bayesian deep learning for computer vision? -
X-MAN: Explaining multiple sources of anomalies in video Stanislaw ...
mi.eng.cam.ac.uk/~cipolla/archive/Publications/inproceedings/2021-CVPR-XMAN-anomaly-detection.pdf9 Apr 2022: 26], sometimes aug-mented with memory modules [19], and/or optical-flow im-ages [13, 20, 21]. ... Learning memory-guided nor-mality for anomaly detection. In CVPR, 2020. [20] M. -
Paper8-features-matching.dvi
mi.eng.cam.ac.uk/~cipolla/lectures/PartIB/old/2022-IB-Paper8-CV-Features-Matrching.pdf26 Apr 2022: Sig. nal. Sigma = 20. As σ increases, the signal is smoothed more and more, and. ... description. Original Level 0. σ = 5. σ = 10. Level 1 Level 2σ = 20. -
Real-time analogue gauge transcription on mobile phone Ben…
mi.eng.cam.ac.uk/~cipolla/archive/Publications/inproceedings/2021-CVPR-analogue-meter-reading.pdf9 Apr 2022: meter_e. 0. 100. 200. test2. 0 20 40 60 80 100 120 140Frame no. ... 4. [20] R. Sablatnig and W. G. Kropatsch. Automatic reading ofanalog display instruments. -
SENGUPTA ET AL.: PROBABILISTIC HUMAN SHAPE & POSE WITH ...
mi.eng.cam.ac.uk/~cipolla/publications/inproceedings/2021-BMVC-body-measurement-reconstruction.pdf9 Apr 2022: to this task yield impressive human pose estimates[6, 9, 10, 15, 20, 21, 25, 45]. ... 36] - - - 24.4 20.6 20.4 15.2 13.6 13.3 90.9 61.0VIBE [19] - - - - 50.1 - - 24.1 - - 51.9. -
FIERY: Future Instance Prediction in Bird’s-Eye Viewfrom Surround…
mi.eng.cam.ac.uk/~cipolla/publications/inproceedings/2021-ICCV-Future-Instance-Prediction-BEV.pdf9 Apr 2022: NuScenes contains 1000 scenes, each 20 sec-onds in length, annotated at 2Hz. ... 5.3. Analysis. 20 22 24 26 28 30. FIERY. Deterministic. Uniform depth. -
Hierarchical Kinematic Probability Distributions for 3D Human Shape…
mi.eng.cam.ac.uk/~cipolla/publications/inproceedings/2021-ICCV-3D-human-shape-in-wild.pdf9 Apr 2022: 3] - 55.6DaNet [63] 82.4 54.8HMR (unpaired) [20] 126.3 92.0Kundu et al. ... Figure 4(a) shows. Max. inputset size. Method SSP-3DPVE-T-SC. HMR [20] 22.9GraphCMR [27] 19.5. -
R. MECCA ET. AL : LUCES: NEAR-FIELD PHOTOMETRIC STEREO ...
mi.eng.cam.ac.uk/~cipolla/archive/Publications/inproceedings/2021-BMVC-LUCES-photometric-stereo-dataset.pdf9 Apr 2022: Xiong et al. [41] have proposed a dataset of 7 objects using 20 directionallights calibrated with two chrome spheres. ... 4 ExperimentsIn this section, we evaluate four competing near-field methods namely [18, 20, 32, 35]. -
Estimating and Exploiting the Aleatoric Uncertaintyin Surface Normal…
mi.eng.cam.ac.uk/~cipolla/archive/Publications/inproceedings/2021-ICCV-surface-normal-uncertainty.pdf9 Apr 2022: We assumethat the uncertainty is heteroscedastic [20] (i.e. certain pix-els have higher uncertainty than the others). ... 5. [20] Alex Kendall and Yarin Gal. What uncertainties do weneed in bayesian deep learning for computer vision? -
FIERY: Future Instance Prediction in Bird’s-Eye Viewfrom Surround…
mi.eng.cam.ac.uk/~cipolla/publications/inproceedings/2021-FIERY-future-instance-BEV.pdf9 Apr 2022: NuScenes contains 1000 scenes, each 20 sec-onds in length, annotated at 2Hz. ... 20 22 24 26 28 30. FIERY. Deterministic. Uniform depth. No future flow. -
Part IA Computing CourseTutorial Guide to C++ Programming Roberto ...
mi.eng.cam.ac.uk/~cipolla/resource/tutorial.pdf12 Apr 2022: 195.2 Input of data from the keyboard using input stream. 20. ... 20. Part IA Computing Course Session 1B. A. Objectives. After reading through sections 3 to 5 of the tutorial guide and working through the ex-amples you should be able to:. -
PX-NET: Simple and Efficient Pixel-Wise Trainingof Photometric Stereo …
mi.eng.cam.ac.uk/~cipolla/publications/inproceedings/2021-ICCV-PX-NET-photometric-normals.pdf9 Apr 2022: Startingfrom the basic linear light response for diffuse reflection[21, 13], more specular behaviour of reflected light havebeen proposed [31, 3, 8, 20, 38, 40]. ... Test-timeaccuracy evolution of CNN-PS [16] network when trained. in total on 20, 30, 40 -
SENGUPTA ET AL.: PROBABILISTIC HUMAN SHAPE & POSE WITH ...
mi.eng.cam.ac.uk/~cipolla/archive/Publications/inproceedings/2021-BMVC-body-measurement-reconstruction.pdf9 Apr 2022: to this task yield impressive human pose estimates[6, 9, 10, 15, 20, 21, 25, 45]. ... 36] - - - 24.4 20.6 20.4 15.2 13.6 13.3 90.9 61.0VIBE [19] - - - - 50.1 - - 24.1 - - 51.9. -
Hierarchical Kinematic Probability Distributions for 3D Human Shape…
mi.eng.cam.ac.uk/~cipolla/archive/Publications/inproceedings/2021-ICCV-3D-human-shape-in-wild.pdf9 Apr 2022: 3] - 55.6DaNet [63] 82.4 54.8HMR (unpaired) [20] 126.3 92.0Kundu et al. ... Figure 4(a) shows. Max. inputset size. Method SSP-3DPVE-T-SC. HMR [20] 22.9GraphCMR [27] 19.5. -
Lifted Semantic Graph Embedding for Omnidirectional Place Recognition
mi.eng.cam.ac.uk/~cipolla/publications/inproceedings/2021-3DV-omnidirectional-localisation.pdf9 Apr 2022: ing [5, 4]. State-of-the-art methods which finetune networkend-to-end for place recognition include NetVLAD [1] forVLAD [20] and [28] for Fisher Vector [29]. ... 4321. [20] Hervé Jégou, Matthijs Douze, Cordelia Schmid, and PatrickPérez. Aggregating -
PX-NET: Simple and Efficient Pixel-Wise Trainingof Photometric Stereo …
mi.eng.cam.ac.uk/~cipolla/archive/Publications/inproceedings/2021-ICCV-PX-NET-photometric-normals.pdf9 Apr 2022: Startingfrom the basic linear light response for diffuse reflection[21, 13], more specular behaviour of reflected light havebeen proposed [31, 3, 8, 20, 38, 40]. ... Test-timeaccuracy evolution of CNN-PS [16] network when trained. in total on 20, 30, 40 -
ACCEPTED FOR PUBLICATION IN IEEE/ACM TRANSACTIONS ON AUDIO, SPEECH,…
mi.eng.cam.ac.uk/~mjfg/ALTA/publications/IEEEACMTransASLP2022_Ragni_Confidence.pdf11 Apr 2022: P(Crefj |Ci, O) =. wrefj Crefj. P(wrefj |Ci, O)P(wrefj |Crefj , O) (20). ... Word error rates forthose languages commonly range between 20-60% [72] andnecessitate the use of error mitigation approaches, such asconfidence scores, to achieve high -
Probabilistic 3D Human Shape and Pose Estimation From Multiple…
mi.eng.cam.ac.uk/~cipolla/archive/Publications/inproceedings/2021-CVPR-3D-body-shape-in-wild.pdf9 Apr 2022: 57, 36, 45, 38, 41, 50], ii) video [26, 20, 47, 40, 16] with. ... methods [20, 26, 47, 49, 40] modify single-image predictors. to take sequences of frames as inputs. -
Lifted Semantic Graph Embedding for Omnidirectional Place Recognition
mi.eng.cam.ac.uk/~cipolla/archive/Publications/inproceedings/2021-3DV-omnidirectional-localisation.pdf9 Apr 2022: ing [5, 4]. State-of-the-art methods which finetune networkend-to-end for place recognition include NetVLAD [1] forVLAD [20] and [28] for Fisher Vector [29]. ... 4321. [20] Hervé Jégou, Matthijs Douze, Cordelia Schmid, and PatrickPérez. Aggregating -
Vision Encoders in Visual Question Answering
mi.eng.cam.ac.uk/~wjb31/Ryan_Anderson_Vision_Encoders_in_VQA.pdf10 Sep 2022: Our results show that explicit alignment enables our VLMs to achieve a significantly higherzero-shot (34.49% vs 20.89%) and best overall (40.39% vs 30.83%) VQA score on ... 4.1 Architecture. 19. 4.1.1 Frozen pretrained LM. 20. 4.1.2 Frozen pretrained -
Improving Attention-based Sequence-to-sequence Models
mi.eng.cam.ac.uk/~mjfg/thesis_qd212.pdf5 Jul 2022: 20. 3.1 Illustration of an encoder-decoder model without attention [168]. <BOS>and <EOS> are special tokens for the beginning and the end of thesequence. ... Equations 2.16 and 2.17 become. cl =Ll′=1. αl,l′hl′ (2.20). αl,l′ =exp(f (hl, hl′ ; -
4F12-examples-2.dvi
mi.eng.cam.ac.uk/~cipolla/lectures/4F12/Examples/4F12-examples-2.pdf17 Oct 2022: The camerais calibrated by observing the image of three markers placed 0, 20 and 30m alongthe track. -
Towards Learning Orientated Assessment for Non-native Learner Spoken…
mi.eng.cam.ac.uk/~mjfg/ALTA/presentations/ALTA_Sheffield_20190306.pdf21 Feb 2022: 400 hour BULATS training set. 20. AM LM % WER. -
Use of Deep Learning in Free Speaking Non-native English Assessment
mi.eng.cam.ac.uk/~mjfg/ALTA/presentations/TSD2021_Knill.pdf21 Feb 2022: dependent. • General approach tunable approach based on deep learning. 20/56. Model-based Pronunciation Features. ... General approach tunable approach based on deep learning. 20/56. Deep Learning Pronunciation Features [5]. -
Applying Deep Learning in Non-native Spoken English Assessment
mi.eng.cam.ac.uk/~mjfg/ALTA/presentations/APSIPA2019_Knill.pdf21 Feb 2022: 20/45. Assessment: Gaussian Process [14, 16]. • Gaussian process• non-parametric model based on joint-Gaussian assumption. • ... 16-20, 2017, 2017, pp. -
Part IA Computing CourseLent Term Software Design Exercise Roberto ...
mi.eng.cam.ac.uk/~cipolla/resource/lent.pdf12 Apr 2022: and balance arrays (day 0) to be 20.0 and 20000.0 respectively. ... 20. 8 Notes on Implementation of Functions of Part II. 8.1 Electronic trading library functions. -
ENGINEERING TRIPOS PART IIB ELECTRICAL AND INFORMATION SCIENCES…
mi.eng.cam.ac.uk/~cipolla/resource/4F12exam.pdf12 Apr 2022: 20%]. (ii) Give an expression for computing the intensity of a smoothed pixel. ... 20%]. (c) Outline an algorithm to recover the elements of the projection matrix. -
Paper8-CV-intro.dvi
mi.eng.cam.ac.uk/~cipolla/lectures/PartIB/old/2022-IB-Paper8-CV-Introduction.pdf26 Apr 2022: 20 Engineering Part IB: Paper 8 Information Engineering. Syllabus. 1. Introduction. • -
FIERY: Future Instance Prediction in Bird’s-Eye Viewfrom Surround…
mi.eng.cam.ac.uk/~cipolla/archive/Publications/inproceedings/2021-FIERY-future-instance-BEV.pdf9 Apr 2022: NuScenes contains 1000 scenes, each 20 sec-onds in length, annotated at 2Hz. ... 20 22 24 26 28 30. FIERY. Deterministic. Uniform depth. No future flow. -
Scaling digital screen reading with one-shot learning and…
mi.eng.cam.ac.uk/~cipolla/publications/inproceedings/2021-WACV-scaling-digital-meters.pdf9 Apr 2022: 0. 20. 40. 60. 80. 100. Prec. isio. n (%. ). Multimeter. 0 5 10 15 20 25Norm dist from GT (px). ... 20] E. Rublee, V. Rabaud, K. Konolige, and G. Bradski. Orb: Anefficient alternative to sift or surf. -
solutions2.dvi
mi.eng.cam.ac.uk/~cipolla/lectures/4F12/Examples/solutions/4F12-examples-2-solutions.pdf17 Oct 2022: X=0, y=0. 0 =sy. s= p22. X=20, y=0.5. 0.5 =sy. s=. ... sys. ]. =. [. 1 01 20. ][. X1. ]. With this calibration we can recover structure X given the y component of the imageposition:. -
X-MAN: Explaining multiple sources of anomalies in video Stanislaw ...
mi.eng.cam.ac.uk/~cipolla/publications/inproceedings/2021-CVPR-XMAN-anomaly-detection.pdf9 Apr 2022: 26], sometimes aug-mented with memory modules [19], and/or optical-flow im-ages [13, 20, 21]. ... Learning memory-guided nor-mality for anomaly detection. In CVPR, 2020. [20] M. -
Real-time analogue gauge transcription on mobile phone Ben…
mi.eng.cam.ac.uk/~cipolla/publications/inproceedings/2021-CVPR-analogue-meter-reading.pdf9 Apr 2022: meter_e. 0. 100. 200. test2. 0 20 40 60 80 100 120 140Frame no. ... 4. [20] R. Sablatnig and W. G. Kropatsch. Automatic reading ofanalog display instruments. -
FIERY: Future Instance Prediction in Bird’s-Eye Viewfrom Surround…
mi.eng.cam.ac.uk/~cipolla/archive/Publications/inproceedings/2021-ICCV-Future-Instance-Prediction-BEV.pdf9 Apr 2022: NuScenes contains 1000 scenes, each 20 sec-onds in length, annotated at 2Hz. ... 5.3. Analysis. 20 22 24 26 28 30. FIERY. Deterministic. Uniform depth. -
Probabilistic 3D Human Shape and Pose Estimation From Multiple…
mi.eng.cam.ac.uk/~cipolla/publications/inproceedings/2021-CVPR-3D-body-shape-in-wild.pdf9 Apr 2022: 57, 36, 45, 38, 41, 50], ii) video [26, 20, 47, 40, 16] with. ... methods [20, 26, 47, 49, 40] modify single-image predictors. to take sequences of frames as inputs.
Search history
Recently clicked results
Recently clicked results
Your click history is empty.
Recent searches
Recent searches
Your search history is empty.