CVPR CVPR

#**** #****
CVPR 2018 Submission #****. CONFIDENTIAL REVIEW COPY. DO NOT DISTRIBUTE.

000 054
001 055
002 056
003 LATEX Author Guidelines for CVPR Proceedings 057
004 058
005 059
006 060
Anonymous CVPR submission
007 061
008 062
009 Paper ID **** 063
010 064
011 065
012 066
Abstract such papers will not be reviewed is that there is no provi-
013 067
sion for supervised revisions of manuscripts. The review-
014 068
The ABSTRACT is to be in fully-justified italicized text, ing process cannot determine the suitability of the paper for
015 069
at the top of the left-hand column, below the author and presentation in eight pages if it is reviewed in eleven.
016 070
affiliation information. Use the word “Abstract” as the title,
017
in 12-point Times, boldface type, centered relative to the 1.4. The ruler 071
018 072
column, initially capitalized. The abstract is to be in 10- The LATEX style defines a printed ruler which should be
019 073
point, single-spaced type. Leave two blank lines after the present in the version submitted for review. The ruler is
020 074
Abstract, then begin the main text. Look at previous CVPR provided in order that reviewers may comment on partic-
021 075
abstracts to get a feel for style and length. ular lines in the paper without circumlocution. If you are
022 076
preparing a document using a non-LATEX document prepa-
023 077
ration system, please arrange for an equivalent ruler to
024 078
1. Introduction appear on the final output pages. The presence or ab-
025 079
sence of the ruler should not change the appearance of
026 Please follow the steps outlined below when submitting 080
any other content on the page. The camera ready copy
027 your manuscript to the IEEE Computer Society Press. This 081
should not contain a ruler. (LATEX users may uncomment
028 style guide now has several important modifications (for ex- 082
the \cvprfinalcopy command in the document pream-
029 ample, you are no longer warned against the use of sticky 083
ble.) Reviewers: note that the ruler measurements do not
030 tape to attach your artwork to the paper), so all authors 084
align well with lines in the paper — this turns out to be very
031 should read this new version. 085
difficult to do well when the paper contains many figures
032 086
1.1. Language and equations, and, when done, looks ugly. Just use frac-
033 087
tional references (e.g. this line is 095.5), although in most
034 088
All manuscripts must be in English. cases one would expect that the approximate location will
035 089
be adequate.
036 1.2. Dual submission 090
037 1.5. Mathematics 091
038 Please refer to the author guidelines on the CVPR 2018 092
039 web page for a discussion of the policy on dual submissions. Please number all of your sections and displayed equa- 093
040 tions. It is important for readers to be able to refer to any 094
1.3. Paper length particular equation. Just because you didn’t refer to it in
041 095
042 Papers, excluding the references section, must be no the text doesn’t mean some future reader might not need 096
043 longer than eight pages in length. The references section to refer to it. It is cumbersome to have to use circumlo- 097
044 will not be included in the page count, and there is no limit cutions like “the equation second from the top of page 3 098
045 on the length of the references section. For example, a pa- column 1”. (Note that the ruler will not be present in the 099
046 per of eight pages with two pages of references would have final copy, so is not an alternative to equation numbers). 100
047 a total length of 10 pages. There will be no extra page All authors will benefit from reading Mermin’s description 101
048 charges for CVPR 2018. of how to write mathematics: http://www.pamitc. 102
049 Overlength papers will simply not be reviewed. This in- org/documents/mermin.pdf. 103
050 cludes papers where the margins and formatting are deemed 104
1.6. Blind review
051 to have been significantly altered from those laid down by 105
052 this style guide. Note that this LATEX guide already sets fig- Many authors misunderstand the concept of anonymiz- 106
053 ure captions and references in a smaller font. The reason ing for blind review. Blind review does not mean that one 107

1
CVPR CVPR
#**** #****
CVPR 2018 Submission #****. CONFIDENTIAL REVIEW COPY. DO NOT DISTRIBUTE.

108 162
must remove citations to one’s own work—in fact it is often Sometimes your paper is about a problem which you
109 163
impossible to review a paper unless the previous citations tested using a tool which is widely known to be restricted
110 164
are known and available. to a single institution. For example, let’s say it’s 1969, you
111 165
Blind review means that you do not use the words “my” have solved a key problem on the Apollo lander, and you
112 166
or “our” when citing previous work. That is all. (But see believe that the CVPR70 audience would like to hear about
113 167
below for techreports.) your solution. The work is a development of your cele-
114 168
Saying “this builds on the work of Lucy Smith [1]” does brated 1968 paper entitled “Zero-g frobnication: How be-
115 169
not say that you are Lucy Smith; it says that you are building ing the only people in the world with access to the Apollo
116 170
on her work. If you are Smith and Jones, do not say “as we lander source code makes us a wow at parties”, by Zeus et
117 171
show in [7]”, say “as Smith and Jones show in [7]” and at al.
118 172
the end of the paper, include reference 7 as you would any You can handle this paper like any other. Don’t write
119 173
other cited work. “We show how to improve our previous work [Anonymous,
120 174
An example of a bad paper just asking to be rejected: 1968]. This time we tested the algorithm on a lunar lander
121 175
[name of lander removed for blind review]”. That would be
122 An analysis of the frobnicatable foo filter. silly, and would immediately identify the authors. Instead 176
123 177
In this paper we present a performance analysis write the following:
124 178
of our previous paper [1], and show it to be in-
125 We describe a system for zero-g frobnication. 179
ferior to all previously known methods. Why the
126 This system is new because it handles the fol- 180
previous paper was accepted without this analysis
127 lowing cases: A, B. Previous systems [Zeus et al. 181
is beyond me.
128 1968] didn’t handle case B properly. Ours han- 182
129 [1] Removed for blind review dles it by including a foo term in the bar integral. 183
130 ... 184
An example of an acceptable paper:
131 The proposed system was integrated with the 185
132 An analysis of the frobnicatable foo filter. Apollo lunar lander, and went all the way to the 186
133 moon, don’t you know. It displayed the following 187
In this paper we present a performance analysis of
134 behaviours which show how well we solved cases 188
the paper of Smith et al. [1], and show it to be in-
135 A and B: ... 189
ferior to all previously known methods. Why the
136 190
previous paper was accepted without this analysis As you can see, the above text follows standard scientific
137 191
is beyond me. convention, reads better than the first version, and does not
138 192
139 [1] Smith, L and Jones, C. “The frobnicatable explicitly name you as the authors. A reviewer might think 193
140 foo filter, a fundamental contribution to human it likely that the new paper was written by Zeus et al., but 194
141 knowledge”. Nature 381(12), 1-213. cannot make any decision based on that guess. He or she 195
142 would have to be sure that no other authors could have been 196
143 If you are making a submission to another conference at contracted to solve problem B. 197
144 the same time, which covers similar or overlapping mate- FAQ: Are acknowledgements OK? No. Leave them for 198
145 rial, you may need to refer to that submission in order to the final copy. 199
146 explain the differences, just as you would if you had pre- 200
147 viously published related work. In such cases, include the 1.7. Miscellaneous 201
148 anonymized parallel submission [4] as additional material Compare the following: 202
149 and cite it as $conf_a$ confa 203
150
[1] Authors. “The frobnicatable foo filter”, F&G $\mathit{conf}_a$ conf a 204
151
2014 Submission ID 324, Supplied as additional See The TEXbook, p165. 205
152
material fg324.pdf. The space after e.g., meaning “for example”, should not 206
153 be a sentence-ending space. So e.g. is correct, e.g. is not. 207
154 Finally, you may feel you need to tell the reader that The provided \eg macro takes care of this. 208
155 more details can be found elsewhere, and refer them to a When citing a multi-author paper, you may save space 209
156 technical report. For conference submissions, the paper by using “et alia”, shortened to “et al.” (not “et. al.” as 210
157 must stand on its own, and not require the reviewer to go “et” is a complete word.) However, use it only when there 211
158 to a techreport for further details. Thus, you may say in are three or more authors. Thus, the following is correct: 212
159 the body of the paper “further details may be found in [5]”. “ Frobnication has been trendy lately. It was introduced 213
160 Then submit the techreport as additional material. Again, by Alpher [1], and subsequently developed by Alpher and 214
161 you may not assume the reviewers will read this material. Fotheringham-Smythe [2], and Alpher et al. [3].” 215

2
CVPR CVPR
#**** #****
CVPR 2018 Submission #****. CONFIDENTIAL REVIEW COPY. DO NOT DISTRIBUTE.

216 270
not capitalize articles, coordinate conjunctions, or preposi-
217 271
tions (unless the title begins with such a word). Leave two
218 272
blank lines after the title.
219 273
AUTHOR NAME(s) and AFFILIATION(s) are to be
220 274
centered beneath the title and printed in Times 12-point,
221 275
non-boldface type. This information is to be followed by
222 276
two blank lines.
223 277
The ABSTRACT and MAIN TEXT are to be in a two-
224 278
column format.
225 279
MAIN TEXT. Type main text in 10-point Times, single-
226 280
spaced. Do NOT use double-spacing. All paragraphs
227 281
should be indented 1 pica (approx. 1/6 inch or 0.422 cm).
228 282
Make sure your text is fully justified—that is, flush left and
229 283
Figure 1. Example of caption. It is set in Roman so that mathemat- flush right. Please do not place any additional blank lines
230 284
ics (always set in Roman: B sin A = A sin B) may be included between paragraphs.
231 285
without an ugly clash. Figure and table captions should be 9-point Roman type
232 286
as in Figures 1 and 2. Short captions should be centred.
233 287
Callouts should be 9-point Helvetica, non-boldface type.
234 This is incorrect: “... subsequently developed by Alpher 288
Initially capitalize only the first word of section titles and
235 et al. [2] ...” because reference [2] has just two authors. 289
first-, second-, and third-order headings.
236 If you use the \etal macro provided, then you need not 290
FIRST-ORDER HEADINGS. (For example, 1. Intro-
237 worry about double periods when used at the end of a sen- 291
duction) should be Times 12-point boldface, initially cap-
238 tence as in Alpher et al. 292
italized, flush left, with one blank line before, and one blank
239 For this citation style, keep multiple citations in numeri- 293
line after.
240 cal (not chronological) order, so prefer [2, 1, 4] to [1, 2, 4]. 294
SECOND-ORDER HEADINGS. (For example, 1.1.
241 295
Database elements) should be Times 11-point boldface,
242 2. Formatting your paper initially capitalized, flush left, with one blank line before, 296
243 297
All text must be in a two-column format. The total allow- and one after. If you require a third-order heading (we dis-
244 298
able width of the text area is 6 78 inches (17.5 cm) wide by courage it), use 10-point Times, boldface, initially capital-
245 299
8 87 inches (22.54 cm) high. Columns are to be 3 14 inches ized, flush left, preceded by one blank line, followed by a
246 300
(8.25 cm) wide, with a 16 5
inch (0.8 cm) space between period and your text on the same line.
247 301
248 them. The main title (on the first page) should begin 1.0 2.3. Footnotes 302
249 inch (2.54 cm) from the top edge of the page. The second 303
and following pages should begin 1.0 inch (2.54 cm) from Please use footnotes1 sparingly. Indeed, try to avoid
250 304
the top edge. On all pages, the bottom margin should be footnotes altogether and include necessary peripheral ob-
251 305
1-1/8 inches (2.86 cm) from the bottom edge of the page servations in the text (within parentheses, if you prefer, as
252 306
for 8.5 × 11-inch paper; for A4 paper, approximately 1-5/8 in this sentence). If you wish to use a footnote, place it at the
253 307
inches (4.13 cm) from the bottom edge of the page. bottom of the column on the page on which it is referenced.
254 308
Use Times 8-point type, single-spaced.
255 309
2.1. Margins and page numbering
256 2.4. References 310
257 All printed material, including text, illustrations, and 311
258 charts, must be kept within a print area 6-7/8 inches (17.5 List and number all bibliographical references in 9-point 312
259 cm) wide by 8-7/8 inches (22.54 cm) high. Times, single-spaced, at the end of your paper. When ref- 313
260
erenced in the text, enclose the citation number in square 314
2.2. Type-style and fonts brackets, for example [4]. Where appropriate, include the
261 315
262 Wherever Times is specified, Times Roman may also be name(s) of editors of referenced books. 316
263 used. If neither is available on your word processor, please 2.5. Illustrations, graphs, and photographs 317
264 use the font closest in appearance to Times to which you 318
265 have access. All graphics should be centered. Please ensure that any 319
266 MAIN TITLE. Center the title 1-3/8 inches (3.49 cm) point you wish to make is resolvable in a printed copy of 320
267 from the top edge of the first page. The title should be the paper. Resize fonts in figures to match the font in the 321
268 in Times 14-point, boldface type. Capitalize the first let- 1 This is what a footnote looks like. It often distracts the reader from 322
269 ter of nouns, pronouns, verbs, adjectives, and adverbs; do the main flow of the argument. 323

3
CVPR CVPR
#**** #****
CVPR 2018 Submission #****. CONFIDENTIAL REVIEW COPY. DO NOT DISTRIBUTE.

324 378
325 379
326 380
327 381
328 382
329 383
330 384
331 385
332 386
333 387
334 388
335 389
336 390
337 391
338 Figure 2. Example of a short caption, which should be centered. 392
339 393
340 394
Method Frobnability References
341 395
Theirs Frumpy [1] A. Alpher. Frobnication. Journal of Foo, 12(1):234–778,
342 396
Yours Frobbly 2002. 2, 3
343 397
Ours Makes one’s heart Frob [2] A. Alpher and J. P. N. Fotheringham-Smythe. Frobnication
344 398
345 revisited. Journal of Foo, 13(1):234–778, 2003. 2, 3 399
Table 1. Results. Ours is better.
346 [3] A. Alpher, J. P. N. Fotheringham-Smythe, and G. Gamow. Can 400
347
a machine frobnicate? Journal of Foo, 14(1):234–778, 2004. 401
2
348 body text, and choose line widths which render effectively 402
[4] Authors. The frobnicatable foo filter, 2014. Face and
349 in print. Many readers (and reviewers), even of an electronic 403
Gesture submission ID 324. Supplied as additional material
350 copy, will choose to print your paper in order to read it. 404
fg324.pdf. 2, 3
351 You cannot insist that they do otherwise, and therefore must 405
[5] Authors. Frobnication tutorial, 2014. Supplied as additional
352 not assume that they can zoom in to see tiny details on a material tr.pdf. 2 406
353 graphic. 407
354 When placing figures in LATEX, it’s almost always best to 408
355 use \includegraphics, and to specify the figure width 409
356 as a multiple of the line width as in the example below 410
357 411
358 412
\usepackage[dvips]{graphicx} ...
359 413
\includegraphics[width=0.8\linewidth]
360 414
{myfile.eps}
361 415
362 416
363 2.6. Color 417
364 418
Please refer to the author guidelines on the CVPR 2018
365 419
web page for a discussion of the use of color in your docu-
366 420
ment.
367 421
368 422
369 3. Final copy 423
370 424
371 You must include your signed IEEE copyright release 425
372 form when you submit your finished paper. We MUST have 426
373 this form before your paper can be published in the proceed- 427
374 ings. 428
375 Please direct any questions to the production editor in 429
376 charge of these proceedings at the IEEE Computer Society 430
377 Press: Phone (714) 821-8380, or Fax (714) 761-1784. 431