forked from jonbarron/website
-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
executable file
·2478 lines (2358 loc) · 136 KB
/
index.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!DOCTYPE HTML>
<html lang="en"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<title>Jon Barron</title>
<meta name="author" content="Jon Barron">
<meta name="viewport" content="width=device-width, initial-scale=1">
<link rel="stylesheet" type="text/css" href="stylesheet.css">
<link rel="icon" type="image/png" href="images/seal_icon.png">
</head>
<body>
<table style="width:100%;max-width:800px;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr style="padding:0px">
<td style="padding:0px">
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr style="padding:0px">
<td style="padding:2.5%;width:63%;vertical-align:middle">
<p style="text-align:center">
<name>Jon Barron</name>
</p>
<p>I am a staff research scientist at <a href="https://ai.google/research">Google Research</a>, where I work on computer vision and machine learning.
</p>
<p>
At Google I've worked on <a href="https://ai.googleblog.com/2020/12/portrait-light-enhancing-portrait.html">Portrait Light</a>, <a href="https://ai.googleblog.com/2014/04/lens-blur-in-new-google-camera-app.html">Lens Blur</a>, <a href="https://ai.googleblog.com/2014/10/hdr-low-light-and-high-dynamic-range.html">HDR+</a>, <a href="https://www.google.com/get/cardboard/jump/">Jump</a>, <a href="https://ai.googleblog.com/2017/10/portrait-mode-on-pixel-2-and-pixel-2-xl.html">Portrait Mode</a>, and <a href="https://www.youtube.com/watch?v=JSnB06um5r4">Glass</a>. I did my PhD at <a href="http://www.eecs.berkeley.edu/">UC Berkeley</a>, where I was advised by <a href="http://www.cs.berkeley.edu/~malik/">Jitendra Malik</a> and funded by the <a href="http://www.nsfgrfp.org/">NSF GRFP</a>. I did my bachelors at the <a href="http://cs.toronto.edu">University of Toronto</a>.
I've received the <a href="https://www2.eecs.berkeley.edu/Students/Awards/15/">C.V. Ramamoorthy Distinguished Research Award</a> and the <a href="https://www.thecvf.com/?page_id=413#YRA">PAMI Young Researcher Award</a>.
</p>
<p style="text-align:center">
<a href="mailto:jonbarron@gmail.com">Email</a>  / 
<a href="data/JonBarron-CV.pdf">CV</a>  / 
<a href="data/JonBarron-bio.txt">Bio</a>  / 
<a href="https://scholar.google.com/citations?hl=en&user=jktWnL8AAAAJ">Google Scholar</a>  / 
<a href="https://twitter.com/jon_barron">Twitter</a>  / 
<a href="https://github.com/jonbarron/">Github</a>
</p>
</td>
<td style="padding:2.5%;width:40%;max-width:40%">
<a href="images/JonBarron.jpg"><img style="width:100%;max-width:100%" alt="profile photo" src="images/JonBarron_circle.jpg" class="hoverZoomLink"></a>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:20px;width:100%;vertical-align:middle">
<heading>Research</heading>
<p>
I'm interested in computer vision, machine learning, optimization, and image processing.
Much of my research is about inferring the physical world (shape, motion, color, light, etc) from images.
Representative papers are <span class="highlight">highlighted</span>.
</p>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="inerf_stop()" onmouseover="inerf_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='inerf_image'><video width=100% height=100% muted autoplay loop>
<source src="images/inerf_after.mp4" type="video/mp4">
Your browser does not support the video tag.
</video></div>
<img src='images/inerf_before.jpg' width="160">
</div>
<script type="text/javascript">
function inerf_start() {
document.getElementById('inerf_image').style.opacity = "1";
}
function inerf_stop() {
document.getElementById('inerf_image').style.opacity = "0";
}
inerf_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="http://yenchenlin.me/inerf/">
<papertitle>iNeRF: Inverting Neural Radiance Fields for Pose Estimation</papertitle>
</a>
<br>
<a href="https://yenchenlin.me/">Lin Yen-Chen</a>,
<a href="http://www.peteflorence.com/">Pete Florence</a>,
<strong>Jonathan T. Barron</strong>,
<a href="https://meche.mit.edu/people/faculty/ALBERTOR@MIT.EDU">Alberto Rodriguez</a>, <br>
<a href="http://web.mit.edu/phillipi/">Phillip Isola</a>,
<a href="https://scholar.google.com/citations?user=_BPdgV0AAAAJ&hl=en">Tsung-Yi Lin</a>
<br>
<em>arXiv</em>, 2020
<br>
<a href="http://yenchenlin.me/inerf/">project page</a> /
<a href="https://arxiv.org/abs/2012.05877">arXiv</a> /
<a href="https://www.youtube.com/watch?v=eQuCZaQN0tI">video</a>
<p></p>
<p>Given an image of an object and a NeRF of that object, you can estimate that object's pose.
</p>
</td>
</tr>
<tr onmouseout="nerd_stop()" onmouseover="nerd_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='nerd_image'><video width=100% height=100% muted autoplay loop>
<source src="images/nerd_160.mp4" type="video/mp4">
Your browser does not support the video tag.
</video></div>
<img src='images/nerd_160.jpg' width="160">
</div>
<script type="text/javascript">
function nerd_start() {
document.getElementById('nerd_image').style.opacity = "1";
}
function nerd_stop() {
document.getElementById('nerd_image').style.opacity = "0";
}
nerd_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://markboss.me/publication/2021-nerd/">
<papertitle>NeRD: Neural Reflectance Decomposition from Image Collections</papertitle>
</a>
<br>
<a href="https://markboss.me">Mark Boss</a>,
<a href="https://uni-tuebingen.de/en/fakultaeten/mathematisch-naturwissenschaftliche-fakultaet/fachbereiche/informatik/lehrstuehle/computergrafik/lehrstuhl/mitarbeiter/raphael-braun/">Raphael Braun</a>,
<a href="https://varunjampani.github.io">Varun Jampani</a>,
<strong>Jonathan T. Barron</strong>,
<a href="http://people.csail.mit.edu/celiu/">Ce Liu</a>,
<a href="https://uni-tuebingen.de/en/faculties/faculty-of-science/departments/computer-science/lehrstuehle/computergrafik/computer-graphics/staff/prof-dr-ing-hendrik-lensch/">Hendrik P. A. Lensch</a>
<br>
<em>arXiv</em>, 2020
<br>
<a href="https://markboss.me/publication/2021-nerd/">project page</a> /
<a href="https://www.youtube.com/watch?v=JL-qMTXw9VU">video</a> /
<a href="https://github.com/cgtuebingen/NeRD-Neural-Reflectance-Decomposition">code</a> /
<a href="https://arxiv.org/abs/2012.03918">arXiv</a>
<p></p>
<p>
A NeRF-like model that can decompose (and mesh) objects with non-Lambertian reflectances, complex geometry, and unknown illumination.
</p>
</td>
</tr>
<tr onmouseout="nerv_stop()" onmouseover="nerv_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='nerv_image'><video width=100% height=100% muted autoplay loop>
<source src="images/hotdog.mp4" type="video/mp4">
Your browser does not support the video tag.
</video></div>
<img src='images/hotdog.jpg' width="160">
</div>
<script type="text/javascript">
function nerv_start() {
document.getElementById('nerv_image').style.opacity = "1";
}
function nerv_stop() {
document.getElementById('nerv_image').style.opacity = "0";
}
nerv_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://people.eecs.berkeley.edu/~pratul/nerv/">
<papertitle>NeRV: Neural Reflectance and Visibility Fields for Relighting and View Synthesis</papertitle>
</a>
<br>
<a href="https://people.eecs.berkeley.edu/~pratul/">Pratul Srinivasan</a>,
<a href="https://boyangdeng.com/">Boyang Deng</a>,
<a href="https://people.csail.mit.edu/xiuming/">Xiuming Zhang</a>,
<a href="http://matthewtancik.com/">Matthew Tancik</a>,
</br>
<a href="https://people.eecs.berkeley.edu/~bmild/">Ben Mildenhall</a>,
<strong>Jonathan T. Barron</strong>
<br>
<em>arXiv</em>, 2020
<br>
<a href="https://people.eecs.berkeley.edu/~pratul/nerv/">project page</a> /
<a href="https://www.youtube.com/watch?v=4XyDdvhhjVo">video</a> /
<a href="https://arxiv.org/abs/2012.03927">arXiv</a>
<p></p>
<p>Using neural approximations of expensive visibility integrals lets you recover relightable NeRF-like models.</p>
</td>
</tr>
<tr onmouseout="winr_stop()" onmouseover="winr_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='winr_image'><video width=100% height=100% muted autoplay loop>
<source src="images/notre_160.mp4" type="video/mp4">
Your browser does not support the video tag.
</video></div>
<img src='images/notre.jpg' width="160">
</div>
<script type="text/javascript">
function winr_start() {
document.getElementById('winr_image').style.opacity = "1";
}
function winr_stop() {
document.getElementById('winr_image').style.opacity = "0";
}
winr_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="http://www.matthewtancik.com/learnit">
<papertitle>Learned Initializations for Optimizing Coordinate-Based Neural Representations</papertitle>
</a>
<br>
<a href="http://matthewtancik.com/">Matthew Tancik*</a>,
<a href="https://people.eecs.berkeley.edu/~bmild/">Ben Mildenhall*</a>,
<a href="https://www.linkedin.com/in/terrance-wang/">Terrance Wang</a>,
<a href="https://www.linkedin.com/in/divi-schmidt-262044180/">Divi Schmidt</a>,
<a href="https://people.eecs.berkeley.edu/~pratul/">Pratul Srinivasan</a>,
</br>
<strong>Jonathan T. Barron</strong>,
<a href="https://www2.eecs.berkeley.edu/Faculty/Homepages/yirenng.html">Ren Ng</a>
<br>
<em>arXiv</em>, 2020
<br>
<a href="http://www.matthewtancik.com/learnit">project page</a> /
<a href="https://www.youtube.com/watch?v=A-r9itCzcyo">video</a> /
<a href="https://arxiv.org/abs/2012.02189">arXiv</a>
<p></p>
<p>Using meta-learning to find weight initializations for coordinate-based MLPs allows them to converge faster and generalize better.</p>
</td>
</tr>
<tr onmouseout="nerfie_stop()" onmouseover="nerfie_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='nerfie_image'><video width=100% height=100% muted autoplay loop>
<source src="images/nerfie_after.mp4" type="video/mp4">
Your browser does not support the video tag.
</video></div>
<img src='images/nerfie_before.jpg' width="160">
</div>
<script type="text/javascript">
function nerfie_start() {
document.getElementById('nerfie_image').style.opacity = "1";
}
function nerfie_stop() {
document.getElementById('nerfie_image').style.opacity = "0";
}
nerfie_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://nerfies.github.io/">
<papertitle>Deformable Neural Radiance Fields</papertitle>
</a>
<br>
<a href="https://keunhong.com">Keunhong Park</a>,
<a href="https://utkarshsinha.com">Utkarsh Sinha</a>,
<strong>Jonathan T. Barron</strong>,
<a href="http://sofienbouaziz.com">Sofien Bouaziz</a>,
<a href="https://www.danbgoldman.com">Dan B Goldman</a>,
<a href="https://homes.cs.washington.edu/~seitz/">Steven M. Seitz</a>,
<a href="http://www.ricardomartinbrualla.com">Ricardo-Martin Brualla</a>
<br>
<em>arXiv</em>, 2020
<br>
<a href="https://nerfies.github.io/">project page</a> /
<a href="https://arxiv.org/abs/2011.12948">arXiv</a> /
<a href="https://www.youtube.com/watch?v=MrKrnHhk8IA">video</a>
<p></p>
<p>Building deformation fields into NeRF lets you capture non-rigid subjects, like people.
</p>
</td>
</tr>
<tr onmouseout="flare_stop()" onmouseover="flare_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='flare_image'>
<img src='images/flare_after.jpg' width="160"></div>
<img src='images/flare_before.jpg' width="160">
</div>
<script type="text/javascript">
function flare_start() {
document.getElementById('flare_image').style.opacity = "1";
}
function flare_stop() {
document.getElementById('flare_image').style.opacity = "0";
}
flare_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://arxiv.org/abs/2011.12485">
<papertitle>Single-Image Lens Flare Removal</papertitle>
</a>
<br>
<a href="http://yicheng.rice.edu/">Yicheng Wu</a>,
<a href="https://scholar.google.com/citations?user=BxqV_RsAAAAJ">Qiurui He</a>,
<a href="https://people.csail.mit.edu/tfxue/">Tianfan Xue</a>,
<a href="http://rahuldotgarg.appspot.com/">Rahul Garg</a>,
<a href="http://people.csail.mit.edu/jiawen/">Jiawen Chen</a>,
<a href="https://computationalimaging.rice.edu/team/ashok-veeraraghavan/">Ashok Veeraraghavan</a>,
<strong>Jonathan T. Barron</strong>
<br>
<em>arXiv</em>, 2020
<br>
<p></p>
<p>
Simulating the optics of a camera's lens lets you train a model that removes lens flare from a single image.
</p>
</td>
</tr>
<tr onmouseout="c5_stop()" onmouseover="c5_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='c5_image'>
<img src='images/c5_after.jpg' width="160"></div>
<img src='images/c5_before.jpg' width="160">
</div>
<script type="text/javascript">
function c5_start() {
document.getElementById('c5_image').style.opacity = "1";
}
function c5_stop() {
document.getElementById('c5_image').style.opacity = "0";
}
c5_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://arxiv.org/abs/2011.11890">
<papertitle>Cross-Camera Convolutional Color Constancy</papertitle>
</a>
<br>
<a href="https://sites.google.com/corp/view/mafifi">Mahmoud Afifi</a>,
<strong>Jonathan T. Barron</strong>,
<a href="http://www.chloelegendre.com/">Chloe LeGendre</a>,
<a href="https://research.google/people/105312/">Yun-Ta Tsai</a>,
<a href="https://www.linkedin.com/in/fbleibel/">Francois Bleibel</a>
<br>
<em>arXiv</em>, 2020
<br>
<p></p>
<p>
With some extra (unlabeled) test-set images, you can build a hypernetwork that calibrates itself at test time to previously-unseen cameras.
</p>
</td>
</tr>
<tr onmouseout="lssr_stop()" onmouseover="lssr_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='lssr_image'>
<img src='images/lssr_after.jpg' width="160"></div>
<img src='images/lssr_before.jpg' width="160">
</div>
<script type="text/javascript">
function lssr_start() {
document.getElementById('lssr_image').style.opacity = "1";
}
function lssr_stop() {
document.getElementById('lssr_image').style.opacity = "0";
}
lssr_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="http://cseweb.ucsd.edu/~viscomp/projects/SIGA20LightstageSuperres/">
<papertitle>Light Stage Super-Resolution: Continuous High-Frequency Relighting</papertitle>
</a>
<br>
<a href="http://kevinkingo.com/">Tiancheng Sun</a>,
<a href="https://cseweb.ucsd.edu/~zex014/">Zexiang Xu</a>
<a href="http://people.csail.mit.edu/xiuming/">Xiuming Zhang</a>,
<a href="http://www.seanfanello.it/">Sean Fanello</a>,
<a href="https://scholar.google.com/citations?user=5D0_pjcAAAAJ&hl=en">Christoph Rhemann</a>,
<a href="https://www.pauldebevec.com/">Paul Debevec</a>,
<a href="https://research.google/people/105312/">Yun-Ta Tsai</a>,
<strong>Jonathan T. Barron</strong>,
<a href="https://cseweb.ucsd.edu/~ravir/">Ravi Ramamoorthi</a>
<br>
<em>SIGGRAPH Asia</em>, 2020
<br>
<a href="http://cseweb.ucsd.edu/~viscomp/projects/SIGA20LightstageSuperres/">project page</a> /
<a href="https://arxiv.org/abs/2010.08888">arXiv</a>
<p></p>
<p>
Scans for light stages are inherently aliased, but we can use learning to super-resolve them.
</p>
</td>
</tr>
<tr onmouseout="dualrefl_stop()" onmouseover="dualrefl_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='dualrefl_image'>
<img src='images/dualrefl_after.jpg' width="160"></div>
<img src='images/dualrefl_before.jpg' width="160">
</div>
<script type="text/javascript">
function dualrefl_start() {
document.getElementById('dualrefl_image').style.opacity = "1";
}
function dualrefl_stop() {
document.getElementById('dualrefl_image').style.opacity = "0";
}
dualrefl_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="http://sniklaus.com/dualref">
<papertitle>Learned Dual-View Reflection Removal</papertitle>
</a>
<br>
<a href="http://sniklaus.com/welcome">Simon Niklaus</a>,
<a href="https://people.eecs.berkeley.edu/~cecilia77/">Xuaner (Cecilia) Zhang</a>,
<strong>Jonathan T. Barron</strong>,
<a href="http://nealwadhwa.com">Neal Wadhwa</a>,
<a href="http://rahuldotgarg.appspot.com/">Rahul Garg</a>,
<a href="http://web.cecs.pdx.edu/~fliu/">Feng Liu</a>,
<a href="https://people.csail.mit.edu/tfxue/">Tianfan Xue</a>,
<br>
<em>WACV</em>, 2021
<br>
<a href="http://sniklaus.com/dualref">project page</a> /
<a href="https://arxiv.org/abs/2010.00702">arXiv</a>
<p></p>
<p>
Reflections and the things behind them often exhibit parallax, and this lets you remove reflections from stereo pairs.
</p>
</td>
</tr>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='nlt_image'><video width=100% height=100% muted autoplay loop>
<source src="images/nlt_after.mp4" type="video/mp4">
Your browser does not support the video tag.
</video></div>
<img src='images/nlt_before.jpg' width="160">
</div>
<script type="text/javascript">
function nlt_start() {
document.getElementById('nlt_image').style.opacity = "1";
}
function nlt_stop() {
document.getElementById('nlt_image').style.opacity = "0";
}
nlt_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="http://nlt.csail.mit.edu/">
<papertitle>Neural Light Transport for Relighting and View Synthesis</papertitle>
</a>
<br>
<a href="http://people.csail.mit.edu/xiuming/">Xiuming Zhang</a>,
<a href="http://www.seanfanello.it/">Sean Fanello</a>,
<a href="https://research.google/people/105312/">Yun-Ta Tsai</a>,
<a href="http://kevinkingo.com/">Tiancheng Sun</a>,
<a href="https://people.csail.mit.edu/tfxue/">Tianfan Xue</a>,
<a href="https://research.google/people/106687/">Rohit Pandey</a>,
<a href="https://www.dtic.ua.es/~sorts/">Sergio Orts-Escolano</a>,
<a href="https://dl.acm.org/profile/99659224296">Philip Davidson</a>,
<a href="https://scholar.google.com/citations?user=5D0_pjcAAAAJ&hl=en">Christoph Rhemann</a>,
<a href="http://www.pauldebevec.com/">Paul Debevec</a>,
<strong>Jonathan T. Barron</strong>,
<a href="http://cseweb.ucsd.edu/~ravir/">Ravi Ramamoorthi</a>,
<a href="http://billf.mit.edu/">William T. Freeman</a>
<br>
<em>arXiv</em>, 2020
<br>
<a href="http://nlt.csail.mit.edu/">project page</a> /
<a href="https://arxiv.org/abs/2008.03806">arXiv</a> /
<a href="https://www.youtube.com/watch?v=OGEnCWZihHE">video</a>
<p></p>
<p>Embedding a convnet within a predefined texture atlas enables simultaneous view synthesis and relighting.</p>
</td>
</tr>
<tr onmouseout="nerfw_stop()" onmouseover="nerfw_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='nerfw_image'><video width=100% height=100% muted autoplay loop>
<source src="images/nerfw_after.mp4" type="video/mp4">
Your browser does not support the video tag.
</video></div>
<img src='images/nerfw_before.jpg' width="160">
</div>
<script type="text/javascript">
function nerfw_start() {
document.getElementById('nerfw_image').style.opacity = "1";
}
function nerfw_stop() {
document.getElementById('nerfw_image').style.opacity = "0";
}
nerfw_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://nerf-w.github.io/">
<papertitle>NeRF in the Wild: Neural Radiance Fields for Unconstrained Photo Collections</papertitle>
</a>
<br>
<a href="http://www.ricardomartinbrualla.com/">Ricardo Martin-Brualla*</a>,
<a href="https://scholar.google.com/citations?user=g98QcZUAAAAJ&hl=en">Noha Radwan*</a>,
<a href="https://research.google/people/105804/">Mehdi S. M. Sajjadi*</a>,
<strong>Jonathan T. Barron</strong>,
<a href="https://scholar.google.com/citations?user=FXNJRDoAAAAJ&hl=en">Alexey Dosovitskiy</a>,
<a href="http://www.stronglyconvex.com/about.html">Daniel Duckworth</a>
<br>
<em>arXiv</em>, 2020
<br>
<a href="https://nerf-w.github.io/">project page</a> /
<a href="https://arxiv.org/abs/2008.02268">arXiv</a> /
<a href="https://www.youtube.com/watch?v=yPKIxoN2Vf0">video</a>
<p></p>
<p>Letting NeRF reason about occluders and appearance variation produces photorealistic view synthesis using only unstructured internet photos.</p>
</td>
</tr>
<tr onmouseout="ff_stop()" onmouseover="ff_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='ff_image'>
<img src='images/lion_ff.jpg' width="160"></div>
<img src='images/lion_none.jpg' width="160">
</div>
<script type="text/javascript">
function ff_start() {
document.getElementById('ff_image').style.opacity = "1";
}
function ff_stop() {
document.getElementById('ff_image').style.opacity = "0";
}
ff_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://people.eecs.berkeley.edu/~bmild/fourfeat/index.html">
<papertitle>Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains</papertitle>
</a>
<br>
<a href="http://matthewtancik.com/">Matthew Tancik*</a>,
<a href="https://people.eecs.berkeley.edu/~pratul/">Pratul Srinivasan*</a>,
<a href="https://people.eecs.berkeley.edu/~bmild/">Ben Mildenhall*</a>,
<a href="https://people.eecs.berkeley.edu/~sfk/">Sara Fridovich-Keil</a>,
<a href="https://www.linkedin.com/in/nithinraghavan">Nithin Raghavan</a>,
<a href="https://scholar.google.com/citations?user=lvA86MYAAAAJ&hl=en">Utkarsh Singhal</a>,
<a href="http://cseweb.ucsd.edu/~ravir/">Ravi Ramamoorthi</a>,
<strong>Jonathan T. Barron</strong>,
<a href="https://www2.eecs.berkeley.edu/Faculty/Homepages/yirenng.html">Ren Ng</a>
<br>
<em>NeurIPS</em>, 2020   <font color=#FF8080><strong>(Spotlight)</strong></font>
<br>
<a href="https://people.eecs.berkeley.edu/~bmild/fourfeat/">project page</a> /
video: <a href="https://www.youtube.com/watch?v=nVA6K6Sn2S4">3 min</a>, <a href="https://www.youtube.com/watch?v=iKyIJ_EtSkw">10 min</a> /
<a href="https://arxiv.org/abs/2006.10739">arXiv</a> /
<a href="https://github.com/tancik/fourier-feature-networks">code</a>
<p></p>
<p>Composing neural networks with a simple Fourier feature mapping allows them to learn detailed high-frequency functions.</p>
</td>
</tr>
<tr onmouseout="thresh_stop()" onmouseover="thresh_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='thresh_image'>
<img src='images/thresh_after.png' width="160"></div>
<img src='images/thresh_before.jpg' width="160">
</div>
<script type="text/javascript">
function thresh_start() {
document.getElementById('thresh_image').style.opacity = "1";
}
function thresh_stop() {
document.getElementById('thresh_image').style.opacity = "0";
}
thresh_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://arxiv.org/abs/2007.07350">
<papertitle>A Generalization of Otsu's Method and Minimum Error Thresholding</papertitle>
</a>
<br>
<strong>Jonathan T. Barron</strong>
<br>
<em>ECCV</em>, 2020   <font color=#FF8080><strong>(Spotlight)</strong></font>
<br>
<a href="https://github.com/jonbarron/hist_thresh">code</a> /
<a href="https://www.youtube.com/watch?v=rHtQQlQo1Q4">video</a> /
<a href="data/BarronECCV2020.bib">bibtex</a>
<br>
<p></p>
<p>
A simple and fast Bayesian algorithm that can be written in ~10 lines of code outperforms or matches giant CNNs on image binarization, and unifies three classic thresholding algorithms.
</p>
</td>
</tr>
<tr onmouseout="uflow_stop()" onmouseover="uflow_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='uflow_image'>
<img src='images/uflow_after.png' width="160"></div>
<img src='images/uflow_before.jpg' width="160">
</div>
<script type="text/javascript">
function uflow_start() {
document.getElementById('uflow_image').style.opacity = "1";
}
function uflow_stop() {
document.getElementById('uflow_image').style.opacity = "0";
}
uflow_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://arxiv.org/abs/2006.04902">
<papertitle>What Matters in Unsupervised Optical Flow</papertitle>
</a>
<br>
<a href="http://ricojonschkowski.com/">Rico Jonschkowski</a>,
<a href="https://www.linkedin.com/in/austin-charles-stone-1ba33b138/">Austin Stone</a>,
<strong>Jonathan T. Barron</strong>,
<a href="https://research.google/people/ArielGordon/">Ariel Gordon</a>,
<a href="https://www.linkedin.com/in/kurt-konolige/">Kurt Konolige</a>,
<a href="https://research.google/people/AneliaAngelova/">Anelia Angelova</a>
<br>
<em>ECCV</em>, 2020   <font color="red"><strong>(Oral Presentation)</strong></font>
<br>
<a href="https://github.com/google-research/google-research/tree/master/uflow">code</a>
<br>
<p></p>
<p>
Extensive experimentation yields a simple optical flow technique that is trained on only unlabeled videos, but still works as well as supervised techniques.
</p>
</td>
</tr>
<tr onmouseout="nerf_stop()" onmouseover="nerf_start()" bgcolor="#ffffd0">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='nerf_image'><video width=100% height=100% muted autoplay loop>
<source src="images/vase_small.mp4" type="video/mp4">
Your browser does not support the video tag.
</video></div>
<img src='images/vase_still.png' width="160">
</div>
<script type="text/javascript">
function nerf_start() {
document.getElementById('nerf_image').style.opacity = "1";
}
function nerf_stop() {
document.getElementById('nerf_image').style.opacity = "0";
}
nerf_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="http://www.matthewtancik.com/nerf">
<papertitle>NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis</papertitle>
</a>
<br>
<a href="https://people.eecs.berkeley.edu/~bmild/">Ben Mildenhall*</a>,
<a href="https://people.eecs.berkeley.edu/~pratul/">Pratul Srinivasan*</a>,
<a href="http://matthewtancik.com/">Matthew Tancik*</a>,
<strong>Jonathan T. Barron</strong>,
<a href="http://cseweb.ucsd.edu/~ravir/">Ravi Ramamoorthi</a>,
<a href="https://www2.eecs.berkeley.edu/Faculty/Homepages/yirenng.html">Ren Ng</a>
<br>
<em>ECCV</em>, 2020   <font color="red"><strong>(Oral Presentation, Best Paper Honorable Mention)</strong></font>
<br>
<a href="http://www.matthewtancik.com/nerf">project page</a>
/
<a href="https://arxiv.org/abs/2003.08934">arXiv</a>
/
<a href="https://www.youtube.com/watch?v=LRAqeM8EjOo&t">talk video</a>
/
<a href="https://www.youtube.com/watch?v=JuH79E8rdKc">supp video</a>
/
<a href="https://github.com/bmild/nerf">code</a>
<p></p>
<p>
Training a tiny non-convolutional neural network to reproduce a scene using volume rendering achieves photorealistic view synthesis.</p>
</td>
</tr>
<tr onmouseout="porshadmanip_stop()" onmouseover="porshadmanip_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='porshadmanip_image'>
<img src='images/porshadmanip_after.jpg' width="160"></div>
<img src='images/porshadmanip_before.jpg' width="160">
</div>
<script type="text/javascript">
function porshadmanip_start() {
document.getElementById('porshadmanip_image').style.opacity = "1";
}
function porshadmanip_stop() {
document.getElementById('porshadmanip_image').style.opacity = "0";
}
porshadmanip_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://arxiv.org/abs/2005.08925">
<papertitle>Portrait Shadow Manipulation</papertitle>
</a>
<br>
<a href="https://people.eecs.berkeley.edu/~cecilia77/">Xuaner (Cecilia) Zhang</a>,
<strong>Jonathan T. Barron</strong>,
<a href="https://ai.google/research/people/105312/">Yun-Ta Tsai</a>,
<a href="https://www.linkedin.com/in/rohit-pandey-bab10b7a/">Rohit Pandey</a>,
<a href="http://people.csail.mit.edu/xiuming/">Xiuming Zhang</a>,
<a href="http://graphics.stanford.edu/~renng/">Ren Ng</a>,
<a href="http://graphics.stanford.edu/~dejacobs/">David E. Jacobs</a>
<br>
<em>SIGGRAPH</em>, 2020
<br>
<a href="https://people.eecs.berkeley.edu/~cecilia77/project-pages/portrait">project page</a> /
<a href="https://www.youtube.com/watch?v=M_qYTXhzyac">video</a>
<p></p>
<p>Networks can be trained to remove shadows cast on human faces and to soften harsh lighting.</p>
</td>
</tr>
<tr onmouseout="learnaf_stop()" onmouseover="learnaf_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='learnaf_image'>
<img src='images/learnaf_after.jpg' width="160"></div>
<img src='images/learnaf_before.jpg' width="160">
</div>
<script type="text/javascript">
function learnaf_start() {
document.getElementById('learnaf_image').style.opacity = "1";
}
function learnaf_stop() {
document.getElementById('learnaf_image').style.opacity = "0";
}
learnaf_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://arxiv.org/abs/2004.12260">
<papertitle>Learning to Autofocus</papertitle>
</a>
<br>
<a href="">Charles Herrmann</a>,
<a href="">Richard Strong Bowen</a>,
<a href="http://nealwadhwa.com">Neal Wadhwa</a>,
<a href="http://rahuldotgarg.appspot.com/">Rahul Garg</a>,
<a href="https://scholar.google.com/citations?user=BxqV_RsAAAAJ">Qiurui He</a>,
<strong>Jonathan T. Barron</strong>,
<a href="http://www.cs.cornell.edu/~rdz/index.htm">Ramin Zabih</a>
<br>
<em>CVPR</em>, 2020
<br>
<a href="https://arxiv.org/abs/2004.12260">arXiv</a>
<p></p>
<p>Machine learning can be used to train cameras to autofocus (which is not the same problem as "depth from defocus").</p>
</td>
</tr>
<tr onmouseout="lighthouse_stop()" onmouseover="lighthouse_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='lh_image'><video width=100% height=100% muted autoplay loop>
<source src="images/rings_crop.mp4" type="video/mp4">
Your browser does not support the video tag.
</video></div>
<img src='images/rings.png' width="160">
</div>
<script type="text/javascript">
function lighthouse_start() {
document.getElementById('lh_image').style.opacity = "1";
}
function lighthouse_stop() {
document.getElementById('lh_image').style.opacity = "0";
}
lighthouse_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://people.eecs.berkeley.edu/~pratul/lighthouse/">
<papertitle>Lighthouse: Predicting Lighting Volumes for Spatially-Coherent Illumination</papertitle>
</a>
<br>
<a href="https://people.eecs.berkeley.edu/~pratul/">Pratul Srinivasan*</a>,
<a href="https://people.eecs.berkeley.edu/~bmild/">Ben Mildenhall*</a>,
<a href="http://matthewtancik.com/">Matthew Tancik</a>,
<strong>Jonathan T. Barron</strong>,
<a href="https://research.google/people/RichardTucker/">Richard Tucker</a>,
<a href="https://www.cs.cornell.edu/~snavely/">Noah Snavely</a>
<br>
<em>CVPR</em>, 2020
<br>
<a href="https://people.eecs.berkeley.edu/~pratul/lighthouse/">project page</a>
/
<a href="https://github.com/pratulsrinivasan/lighthouse">code</a>
/
<a href="https://arxiv.org/abs/2003.08367">arXiv</a>
/
<a href="https://www.youtube.com/watch?v=KsiZpUFPqIU">video</a>
<p></p>
<p>We predict a volume from an input stereo pair that can be used to calculate incident lighting at any 3D point within a scene.</p>
</td>
</tr>
<tr onmouseout="skyopt_stop()" onmouseover="skyopt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='skyopt_image'>
<img src='images/skyopt_after.jpg' width="160"></div>
<img src='images/skyopt_before.jpg' width="160">
</div>
<script type="text/javascript">
function skyopt_start() {
document.getElementById('skyopt_image').style.opacity = "1";
}
function skyopt_stop() {
document.getElementById('skyopt_image').style.opacity = "0";
}
skyopt_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://arxiv.org/abs/2006.10172">
<papertitle>Sky Optimization: Semantically Aware Image Processing of Skies in Low-Light Photography</papertitle>
</a>
<br>
<a href="https://sites.google.com/corp/view/orly-liba/">Orly Liba</a>,
<a href="https://www.linkedin.com/in/longqicai/en-us">Longqi Cai</a>,
<a href="https://ai.google/research/people/105312/">Yun-Ta Tsai</a>,
<a href="https://research.google/people/EladEban/">Elad Eban</a>,
<a href="https://research.google/people/YairMovshovitzAttias/">Yair Movshovitz-Attias</a>,
<a href="https://scholar.google.com/citations?user=2jXxOYQAAAAJ">Yael Pritch</a>,
<a href="https://www.linkedin.com/in/huizhong-chen-00776432">Huizhong Chen</a>,
<strong>Jonathan T. Barron</strong>
<br>
<em>NTIRE CVPRW</em>, 2020
<br>
<a href="https://google.github.io/sky-optimization/">project page</a>
<p></p>
<p>If you want to photograph the sky, it helps to know where the sky is.</p>
</td>
</tr>
<tr onmouseout="nightsight_stop()" onmouseover="nightsight_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='nightsight_image'><img src='images/nightsight_after.jpg'></div>
<img src='images/nightsight_before.jpg'>
</div>
<script type="text/javascript">
function nightsight_start() {
document.getElementById('nightsight_image').style.opacity = "1";
}
function nightsight_stop() {
document.getElementById('nightsight_image').style.opacity = "0";
}
nightsight_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://arxiv.org/abs/1910.11336">
<papertitle>Handheld Mobile Photography in Very Low Light</papertitle>
</a>
<br>
<a href="https://sites.google.com/site/orlylibaprofessional/">Orly Liba</a>,
<a href="https://scholar.google.com/citations?user=6PhlPWMAAAAJ">Kiran Murthy</a>,
<a href="https://ai.google/research/people/105312/">Yun-Ta Tsai</a>,
<a href="https://www.timothybrooks.com/">Timothy Brooks</a>,
<a href="https://people.csail.mit.edu/tfxue/">Tianfan Xue</a>,
<a href="https://scholar.google.com/citations?user=qgc_jY0AAAAJ">Nikhil Karnad</a>,
<a href="https://scholar.google.com/citations?user=BxqV_RsAAAAJ">Qiurui He</a>,
<strong>Jonathan T. Barron</strong>,
<a href="https://ai.google/research/people/105641/">Dillon Sharlet</a>,
<a href="http://www.geisswerks.com/">Ryan Geiss</a>,
<a href="https://people.csail.mit.edu/hasinoff/">Samuel W. Hasinoff</a>,
<a href="https://scholar.google.com/citations?user=2jXxOYQAAAAJ">Yael Pritch</a>,
<a href="http://graphics.stanford.edu/~levoy/">Marc Levoy</a>
<br>
<em>SIGGRAPH Asia</em>, 2019
<br>
<a href="https://github.com/google/night-sight/tree/master/docs">project page</a>
<br>
<p></p>
<p>By rethinking metering, white balance, and tone mapping, we can take pictures in places too dark for humans to see clearly.</p>
</td>
</tr>
<tr onmouseout="font_stop()" onmouseover="font_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='font_image'><img src='images/font_after.png'></div>
<img src='images/font_before.png'>
</div>
<script type="text/javascript">
function font_start() {
document.getElementById('font_image').style.opacity = "1";
}
function font_stop() {
document.getElementById('font_image').style.opacity = "0";
}
font_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://arxiv.org/abs/1910.00748">
<papertitle>A Deep Factorization of Style and Structure in Fonts</papertitle>
</a>
<br>
<a href="http://www.cs.cmu.edu/~asrivats/">Nikita Srivatsan</a>,
<strong>Jonathan T. Barron</strong>,
<a href="https://people.eecs.berkeley.edu/~klein/">Dan Klein</a>,
<a href="http://cseweb.ucsd.edu/~tberg/">Taylor Berg-Kirkpatrick</a>
<br>
<em>EMNLP</em>, 2019   <font color="red"><strong>(Oral Presentation)</strong></font>
<br>
<p></p>
<p>Variational auto-encoders can be used to disentangle a characters style from its content.</p>
</td>
</tr>
<tr onmouseout="dpzlearn_stop()" onmouseover="dpzlearn_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='dpzlearn_image'><img src='images/dpzlearn_after.jpg'></div>
<img src='images/dpzlearn_before.jpg'>
</div>
<script type="text/javascript">
function dpzlearn_start() {
document.getElementById('dpzlearn_image').style.opacity = "1";
}
function dpzlearn_stop() {
document.getElementById('dpzlearn_image').style.opacity = "0";
}
dpzlearn_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://arxiv.org/abs/1904.05822">
<papertitle>Learning Single Camera Depth Estimation using Dual-Pixels</papertitle>
</a>
<br>
<a href="http://rahuldotgarg.appspot.com/">Rahul Garg</a>,
<a href="http://nealwadhwa.com">Neal Wadhwa</a>,
<a href="">Sameer Ansari,</a>,
<strong>Jonathan T. Barron</strong>
<br>
<em>ICCV</em>, 2019   <font color="red"><strong>(Oral Presentation)</strong></font>
<br>
<a href="https://github.com/google-research/google-research/tree/master/dual_pixels">code</a> /
<a href="data/GargICCV2019.bib">bibtex</a>
<p></p>
<p>Considering the optics of dual-pixel image sensors improves monocular depth estimation techniques.</p>
</td>
</tr>
<tr onmouseout="porlight_stop()" onmouseover="porlight_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='porlight_image'><img src='images/porlight_after.jpg'></div>
<img src='images/porlight_before.jpg'>
</div>
<script type="text/javascript">
function porlight_start() {
document.getElementById('porlight_image').style.opacity = "1";
}