-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
1503 lines (1101 loc) · 59.3 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="zh-CN">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1, maximum-scale=2">
<meta name="theme-color" content="#222">
<meta name="generator" content="Hexo 5.0.2">
<link rel="apple-touch-icon" sizes="180x180" href="/images/apple-touch-icon-next.png">
<link rel="icon" type="image/png" sizes="32x32" href="/images/favicon-32x32-next.png">
<link rel="icon" type="image/png" sizes="16x16" href="/images/favicon-16x16-next.png">
<link rel="mask-icon" href="/images/logo.svg" color="#222">
<link rel="stylesheet" href="/css/main.css">
<link rel="stylesheet" href="/lib/font-awesome/css/all.min.css">
<script id="hexo-configurations">
var NexT = window.NexT || {};
var CONFIG = {"hostname":"lt4hyl.top","root":"/","scheme":"Pisces","version":"7.8.0","exturl":false,"sidebar":{"position":"left","display":"post","padding":18,"offset":12,"onmobile":false},"copycode":{"enable":false,"show_result":false,"style":null},"back2top":{"enable":true,"sidebar":false,"scrollpercent":false},"bookmark":{"enable":false,"color":"#222","save":"auto"},"fancybox":false,"mediumzoom":false,"lazyload":false,"pangu":false,"comments":{"style":"tabs","active":null,"storage":true,"lazyload":false,"nav":null},"algolia":{"hits":{"per_page":10},"labels":{"input_placeholder":"Search for Posts","hits_empty":"We didn't find any results for the search: ${query}","hits_stats":"${hits} results found in ${time} ms"}},"localsearch":{"enable":false,"trigger":"auto","top_n_per_article":1,"unescape":false,"preload":false},"motion":{"enable":true,"async":false,"transition":{"post_block":"fadeIn","post_header":"slideDownIn","post_body":"slideDownIn","coll_header":"slideLeftIn","sidebar":"slideUpIn"}}};
</script>
<meta name="description" content="为人民日益增长的美好生活需要而读书：机器学习、深度学习、c++、python、C#">
<meta property="og:type" content="website">
<meta property="og:title" content="Lancelot的小站">
<meta property="og:url" content="http://lt4hyl.top/index.html">
<meta property="og:site_name" content="Lancelot的小站">
<meta property="og:description" content="为人民日益增长的美好生活需要而读书：机器学习、深度学习、c++、python、C#">
<meta property="og:locale" content="zh_CN">
<meta property="article:author" content="刘涛">
<meta property="article:tag" content="机器学习，深度学习，CNN，卷积神经网络，c++,python,C#">
<meta name="twitter:card" content="summary">
<link rel="canonical" href="http://lt4hyl.top/">
<script id="page-configurations">
// https://hexo.io/docs/variables.html
CONFIG.page = {
sidebar: "",
isHome : true,
isPost : false,
lang : 'zh-CN'
};
</script>
<title>Lancelot的小站</title>
<script>
var _hmt = _hmt || [];
(function() {
var hm = document.createElement("script");
hm.src = "https://hm.baidu.com/hm.js?43b02ed86a701dce319ff86325707452";
var s = document.getElementsByTagName("script")[0];
s.parentNode.insertBefore(hm, s);
})();
</script>
<noscript>
<style>
.use-motion .brand,
.use-motion .menu-item,
.sidebar-inner,
.use-motion .post-block,
.use-motion .pagination,
.use-motion .comments,
.use-motion .post-header,
.use-motion .post-body,
.use-motion .collection-header { opacity: initial; }
.use-motion .site-title,
.use-motion .site-subtitle {
opacity: initial;
top: initial;
}
.use-motion .logo-line-before i { left: initial; }
.use-motion .logo-line-after i { right: initial; }
</style>
</noscript>
</head>
<body itemscope itemtype="http://schema.org/WebPage">
<div class="container use-motion">
<div class="headband"></div>
<header class="header" itemscope itemtype="http://schema.org/WPHeader">
<div class="header-inner"><div class="site-brand-container">
<div class="site-nav-toggle">
<div class="toggle" aria-label="切换导航栏">
<span class="toggle-line toggle-line-first"></span>
<span class="toggle-line toggle-line-middle"></span>
<span class="toggle-line toggle-line-last"></span>
</div>
</div>
<div class="site-meta">
<a href="/" class="brand" rel="start">
<span class="logo-line-before"><i></i></span>
<h1 class="site-title">Lancelot的小站</h1>
<span class="logo-line-after"><i></i></span>
</a>
<p class="site-subtitle" itemprop="description">记录点滴成长：AI所向，吾之所往</p>
</div>
<div class="site-nav-right">
<div class="toggle popup-trigger">
</div>
</div>
</div>
<nav class="site-nav">
<ul id="menu" class="main-menu menu">
<li class="menu-item menu-item-home">
<a href="/" rel="section"><i class="fa fa-home fa-fw"></i>首页</a>
</li>
<li class="menu-item menu-item-about">
<a href="/about/" rel="section"><i class="fa fa-user fa-fw"></i>关于</a>
</li>
<li class="menu-item menu-item-archives">
<a href="/archives/" rel="section"><i class="fa fa-archive fa-fw"></i>归档</a>
</li>
</ul>
</nav>
</div>
</header>
<div class="back-to-top">
<i class="fa fa-arrow-up"></i>
<span>0%</span>
</div>
<main class="main">
<div class="main-inner">
<div class="content-wrap">
<div class="content index posts-expand">
<article itemscope itemtype="http://schema.org/Article" class="post-block" lang="zh-CN">
<link itemprop="mainEntityOfPage" href="http://lt4hyl.top/2020-09-07/%E4%BA%8C%E5%8F%89%E6%A0%91%E5%89%8D%E4%B8%AD%E5%90%8E%E5%BA%8F%E9%81%8D%E5%8E%86%E7%9A%84%E4%B8%89%E7%A7%8D%E5%AE%9E%E7%8E%B0.html">
<span hidden itemprop="author" itemscope itemtype="http://schema.org/Person">
<meta itemprop="image" content="/images/avatar.gif">
<meta itemprop="name" content="刘涛">
<meta itemprop="description" content="为人民日益增长的美好生活需要而读书：机器学习、深度学习、c++、python、C#">
</span>
<span hidden itemprop="publisher" itemscope itemtype="http://schema.org/Organization">
<meta itemprop="name" content="Lancelot的小站">
</span>
<header class="post-header">
<h2 class="post-title" itemprop="name headline">
<a href="/2020-09-07/%E4%BA%8C%E5%8F%89%E6%A0%91%E5%89%8D%E4%B8%AD%E5%90%8E%E5%BA%8F%E9%81%8D%E5%8E%86%E7%9A%84%E4%B8%89%E7%A7%8D%E5%AE%9E%E7%8E%B0.html" class="post-title-link" itemprop="url">二叉树前中后序遍历的三种实现</a>
</h2>
<div class="post-meta">
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar"></i>
</span>
<span class="post-meta-item-text">发表于</span>
<time title="创建时间：2020-09-07 22:36:44 / 修改时间：22:47:48" itemprop="dateCreated datePublished" datetime="2020-09-07T22:36:44+08:00">2020-09-07</time>
</span>
<span id="/2020-09-07/%E4%BA%8C%E5%8F%89%E6%A0%91%E5%89%8D%E4%B8%AD%E5%90%8E%E5%BA%8F%E9%81%8D%E5%8E%86%E7%9A%84%E4%B8%89%E7%A7%8D%E5%AE%9E%E7%8E%B0.html" class="post-meta-item leancloud_visitors" data-flag-title="二叉树前中后序遍历的三种实现" title="阅读次数">
<span class="post-meta-item-icon">
<i class="fa fa-eye"></i>
</span>
<span class="post-meta-item-text">阅读次数：</span>
<span class="leancloud-visitors-count"></span>
</span><br>
<span class="post-meta-item" title="本文字数">
<span class="post-meta-item-icon">
<i class="far fa-file-word"></i>
</span>
<span class="post-meta-item-text">本文字数：</span>
<span>5.6k</span>
</span>
<span class="post-meta-item" title="阅读时长">
<span class="post-meta-item-icon">
<i class="far fa-clock"></i>
</span>
<span class="post-meta-item-text">阅读时长 ≈</span>
<span>5 分钟</span>
</span>
</div>
</header>
<div class="post-body" itemprop="articleBody">
<h1 id="二叉树遍历"><a href="#二叉树遍历" class="headerlink" title="二叉树遍历"></a>二叉树遍历</h1><ul>
<li><strong>前序遍历</strong>：<strong>先访问根节点</strong>，再前序遍历左子树，再前序遍历右子树。</li>
<li><strong>中序遍历</strong>：先中序遍历左子树，<strong>再访问根节点</strong>，再中序遍历右子树。</li>
<li><strong>后序遍历</strong>：先后序遍历左子树，再后序遍历右子树，<strong>再访问根节点</strong>。</li>
</ul>
<h1 id="前序遍历的三种实现"><a href="#前序遍历的三种实现" class="headerlink" title="前序遍历的三种实现"></a>前序遍历的三种实现</h1>
<!--noindex-->
<div class="post-button">
<a class="btn" href="/2020-09-07/%E4%BA%8C%E5%8F%89%E6%A0%91%E5%89%8D%E4%B8%AD%E5%90%8E%E5%BA%8F%E9%81%8D%E5%8E%86%E7%9A%84%E4%B8%89%E7%A7%8D%E5%AE%9E%E7%8E%B0.html#more" rel="contents">
阅读全文 »
</a>
</div>
<!--/noindex-->
</div>
<footer class="post-footer">
<div class="post-eof"></div>
</footer>
</article>
<article itemscope itemtype="http://schema.org/Article" class="post-block" lang="zh-CN">
<link itemprop="mainEntityOfPage" href="http://lt4hyl.top/2020-09-05/%E7%9B%AE%E6%A0%87%E6%A3%80%E6%B5%8B%E4%B8%AD%E7%9A%84anchor.html">
<span hidden itemprop="author" itemscope itemtype="http://schema.org/Person">
<meta itemprop="image" content="/images/avatar.gif">
<meta itemprop="name" content="刘涛">
<meta itemprop="description" content="为人民日益增长的美好生活需要而读书：机器学习、深度学习、c++、python、C#">
</span>
<span hidden itemprop="publisher" itemscope itemtype="http://schema.org/Organization">
<meta itemprop="name" content="Lancelot的小站">
</span>
<header class="post-header">
<h2 class="post-title" itemprop="name headline">
<a href="/2020-09-05/%E7%9B%AE%E6%A0%87%E6%A3%80%E6%B5%8B%E4%B8%AD%E7%9A%84anchor.html" class="post-title-link" itemprop="url">目标检测中的anchor</a>
</h2>
<div class="post-meta">
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar"></i>
</span>
<span class="post-meta-item-text">发表于</span>
<time title="创建时间：2020-09-05 16:59:03 / 修改时间：17:03:42" itemprop="dateCreated datePublished" datetime="2020-09-05T16:59:03+08:00">2020-09-05</time>
</span>
<span id="/2020-09-05/%E7%9B%AE%E6%A0%87%E6%A3%80%E6%B5%8B%E4%B8%AD%E7%9A%84anchor.html" class="post-meta-item leancloud_visitors" data-flag-title="目标检测中的anchor" title="阅读次数">
<span class="post-meta-item-icon">
<i class="fa fa-eye"></i>
</span>
<span class="post-meta-item-text">阅读次数：</span>
<span class="leancloud-visitors-count"></span>
</span><br>
<span class="post-meta-item" title="本文字数">
<span class="post-meta-item-icon">
<i class="far fa-file-word"></i>
</span>
<span class="post-meta-item-text">本文字数：</span>
<span>60</span>
</span>
<span class="post-meta-item" title="阅读时长">
<span class="post-meta-item-icon">
<i class="far fa-clock"></i>
</span>
<span class="post-meta-item-text">阅读时长 ≈</span>
<span>1 分钟</span>
</span>
</div>
</header>
<div class="post-body" itemprop="articleBody">
<p><a target="_blank" rel="noopener" href="https://zhuanlan.zhihu.com/p/150332784">目标检测Anchor的What/Where/When/Why/How</a></p>
<p>看到知乎上一位大佬的分享，写的很棒，特贴上链接于此。</p>
</div>
<footer class="post-footer">
<div class="post-eof"></div>
</footer>
</article>
<article itemscope itemtype="http://schema.org/Article" class="post-block" lang="zh-CN">
<link itemprop="mainEntityOfPage" href="http://lt4hyl.top/2020-09-04/%E7%9B%AE%E6%A0%87%E6%A3%80%E6%B5%8B%E6%A6%82%E8%BF%B0.html">
<span hidden itemprop="author" itemscope itemtype="http://schema.org/Person">
<meta itemprop="image" content="/images/avatar.gif">
<meta itemprop="name" content="刘涛">
<meta itemprop="description" content="为人民日益增长的美好生活需要而读书：机器学习、深度学习、c++、python、C#">
</span>
<span hidden itemprop="publisher" itemscope itemtype="http://schema.org/Organization">
<meta itemprop="name" content="Lancelot的小站">
</span>
<header class="post-header">
<h2 class="post-title" itemprop="name headline">
<a href="/2020-09-04/%E7%9B%AE%E6%A0%87%E6%A3%80%E6%B5%8B%E6%A6%82%E8%BF%B0.html" class="post-title-link" itemprop="url">目标检测概述</a>
</h2>
<div class="post-meta">
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar"></i>
</span>
<span class="post-meta-item-text">发表于</span>
<time title="创建时间：2020-09-04 15:22:59" itemprop="dateCreated datePublished" datetime="2020-09-04T15:22:59+08:00">2020-09-04</time>
</span>
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar-check"></i>
</span>
<span class="post-meta-item-text">更新于</span>
<time title="修改时间：2020-09-05 17:00:55" itemprop="dateModified" datetime="2020-09-05T17:00:55+08:00">2020-09-05</time>
</span>
<span id="/2020-09-04/%E7%9B%AE%E6%A0%87%E6%A3%80%E6%B5%8B%E6%A6%82%E8%BF%B0.html" class="post-meta-item leancloud_visitors" data-flag-title="目标检测概述" title="阅读次数">
<span class="post-meta-item-icon">
<i class="fa fa-eye"></i>
</span>
<span class="post-meta-item-text">阅读次数：</span>
<span class="leancloud-visitors-count"></span>
</span><br>
<span class="post-meta-item" title="本文字数">
<span class="post-meta-item-icon">
<i class="far fa-file-word"></i>
</span>
<span class="post-meta-item-text">本文字数：</span>
<span>10k</span>
</span>
<span class="post-meta-item" title="阅读时长">
<span class="post-meta-item-icon">
<i class="far fa-clock"></i>
</span>
<span class="post-meta-item-text">阅读时长 ≈</span>
<span>9 分钟</span>
</span>
</div>
</header>
<div class="post-body" itemprop="articleBody">
<p>对深度学习在目标检测问题上的发展与应用，可以参照：<a href="http://lt4hyl.top/2020-08-28/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%884%EF%BC%89.html">深度学习笔记（4）</a></p>
<p>目标检测三大神器：Detectron2、mmDetection、SimpleDet，均可在Github上找到。</p>
<h1 id="目标检测问题概述"><a href="#目标检测问题概述" class="headerlink" title="目标检测问题概述"></a>目标检测问题概述</h1><h2 id="目标检测问题的定义"><a href="#目标检测问题的定义" class="headerlink" title="目标检测问题的定义"></a>目标检测问题的定义</h2><p>是在图片中对<strong>可变数量</strong>的目标进行查找和分类。</p>
<p>当前目标检测面临的问题：目标种类与数量问题；目标尺度问题；外在环境干扰问题。</p>
<p>计算机视觉领域中的两大基本任务：<strong>目标检测</strong>和<strong>目标分割</strong>。对于计算机视觉的其它任务往往会依赖于这两个的结果来进行后续的处理。比如说目标跟踪，口罩识别等。</p>
<p>其中目标检测包括了：目标定位(Object Localization)和图像分类(Image Classification)。</p>
<p>目标分割可分为：语义分割(Semantic segmentation)和实例分割(Instance Segmentation)。</p>
<!--noindex-->
<div class="post-button">
<a class="btn" href="/2020-09-04/%E7%9B%AE%E6%A0%87%E6%A3%80%E6%B5%8B%E6%A6%82%E8%BF%B0.html#more" rel="contents">
阅读全文 »
</a>
</div>
<!--/noindex-->
</div>
<footer class="post-footer">
<div class="post-eof"></div>
</footer>
</article>
<article itemscope itemtype="http://schema.org/Article" class="post-block" lang="zh-CN">
<link itemprop="mainEntityOfPage" href="http://lt4hyl.top/2020-09-02/%E8%AF%BB%E4%B9%A6%E7%AC%94%E8%AE%B0-%E5%90%B4%E5%86%9B%E5%85%88%E7%94%9F%E4%B8%93%E5%9C%BA.html">
<span hidden itemprop="author" itemscope itemtype="http://schema.org/Person">
<meta itemprop="image" content="/images/avatar.gif">
<meta itemprop="name" content="刘涛">
<meta itemprop="description" content="为人民日益增长的美好生活需要而读书：机器学习、深度学习、c++、python、C#">
</span>
<span hidden itemprop="publisher" itemscope itemtype="http://schema.org/Organization">
<meta itemprop="name" content="Lancelot的小站">
</span>
<header class="post-header">
<h2 class="post-title" itemprop="name headline">
<a href="/2020-09-02/%E8%AF%BB%E4%B9%A6%E7%AC%94%E8%AE%B0-%E5%90%B4%E5%86%9B%E5%85%88%E7%94%9F%E4%B8%93%E5%9C%BA.html" class="post-title-link" itemprop="url">读书笔记-吴军先生专场</a>
</h2>
<div class="post-meta">
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar"></i>
</span>
<span class="post-meta-item-text">发表于</span>
<time title="创建时间：2020-09-02 15:03:47 / 修改时间：17:03:45" itemprop="dateCreated datePublished" datetime="2020-09-02T15:03:47+08:00">2020-09-02</time>
</span>
<span id="/2020-09-02/%E8%AF%BB%E4%B9%A6%E7%AC%94%E8%AE%B0-%E5%90%B4%E5%86%9B%E5%85%88%E7%94%9F%E4%B8%93%E5%9C%BA.html" class="post-meta-item leancloud_visitors" data-flag-title="读书笔记-吴军先生专场" title="阅读次数">
<span class="post-meta-item-icon">
<i class="fa fa-eye"></i>
</span>
<span class="post-meta-item-text">阅读次数：</span>
<span class="leancloud-visitors-count"></span>
</span><br>
<span class="post-meta-item" title="本文字数">
<span class="post-meta-item-icon">
<i class="far fa-file-word"></i>
</span>
<span class="post-meta-item-text">本文字数：</span>
<span>2.7k</span>
</span>
<span class="post-meta-item" title="阅读时长">
<span class="post-meta-item-icon">
<i class="far fa-clock"></i>
</span>
<span class="post-meta-item-text">阅读时长 ≈</span>
<span>2 分钟</span>
</span>
</div>
</header>
<div class="post-body" itemprop="articleBody">
<p>关于吴军先生的介绍请见下图：</p>
<p><img src="/pictures/wujun1.jpg" alt="wujun1" style="zoom:67%;" /></p>
<!--noindex-->
<div class="post-button">
<a class="btn" href="/2020-09-02/%E8%AF%BB%E4%B9%A6%E7%AC%94%E8%AE%B0-%E5%90%B4%E5%86%9B%E5%85%88%E7%94%9F%E4%B8%93%E5%9C%BA.html#more" rel="contents">
阅读全文 »
</a>
</div>
<!--/noindex-->
</div>
<footer class="post-footer">
<div class="post-eof"></div>
</footer>
</article>
<article itemscope itemtype="http://schema.org/Article" class="post-block" lang="zh-CN">
<link itemprop="mainEntityOfPage" href="http://lt4hyl.top/2020-09-02/%E5%B0%8F%E7%9B%AE%E6%A0%87%E6%A3%80%E6%B5%8B%E8%AE%BA%E6%96%87%E9%98%85%E8%AF%BB.html">
<span hidden itemprop="author" itemscope itemtype="http://schema.org/Person">
<meta itemprop="image" content="/images/avatar.gif">
<meta itemprop="name" content="刘涛">
<meta itemprop="description" content="为人民日益增长的美好生活需要而读书：机器学习、深度学习、c++、python、C#">
</span>
<span hidden itemprop="publisher" itemscope itemtype="http://schema.org/Organization">
<meta itemprop="name" content="Lancelot的小站">
</span>
<header class="post-header">
<h2 class="post-title" itemprop="name headline">
<a href="/2020-09-02/%E5%B0%8F%E7%9B%AE%E6%A0%87%E6%A3%80%E6%B5%8B%E8%AE%BA%E6%96%87%E9%98%85%E8%AF%BB.html" class="post-title-link" itemprop="url">小目标检测论文阅读</a>
</h2>
<div class="post-meta">
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar"></i>
</span>
<span class="post-meta-item-text">发表于</span>
<time title="创建时间：2020-09-02 15:03:27 / 修改时间：16:29:40" itemprop="dateCreated datePublished" datetime="2020-09-02T15:03:27+08:00">2020-09-02</time>
</span>
<span id="/2020-09-02/%E5%B0%8F%E7%9B%AE%E6%A0%87%E6%A3%80%E6%B5%8B%E8%AE%BA%E6%96%87%E9%98%85%E8%AF%BB.html" class="post-meta-item leancloud_visitors" data-flag-title="小目标检测论文阅读" title="阅读次数">
<span class="post-meta-item-icon">
<i class="fa fa-eye"></i>
</span>
<span class="post-meta-item-text">阅读次数：</span>
<span class="leancloud-visitors-count"></span>
</span><br>
<span class="post-meta-item" title="本文字数">
<span class="post-meta-item-icon">
<i class="far fa-file-word"></i>
</span>
<span class="post-meta-item-text">本文字数：</span>
<span>4.2k</span>
</span>
<span class="post-meta-item" title="阅读时长">
<span class="post-meta-item-icon">
<i class="far fa-clock"></i>
</span>
<span class="post-meta-item-text">阅读时长 ≈</span>
<span>4 分钟</span>
</span>
</div>
</header>
<div class="post-body" itemprop="articleBody">
<p><strong>论文原文均可从 <code>arxiv</code> 上检索得到。</strong></p>
<h1 id="论文：An-Analysis-of-Scale-Invariance-in-Object-Detection-–-SNIP"><a href="#论文：An-Analysis-of-Scale-Invariance-in-Object-Detection-–-SNIP" class="headerlink" title="论文：An Analysis of Scale Invariance in Object Detection – SNIP"></a>论文：An Analysis of Scale Invariance in Object Detection – SNIP</h1><p>SNIP算法的提出，主要源自对coco数据集和ImageNet数据集中目标尺寸变化的分析。</p>
<ul>
<li>SNIP论文中开篇说到：COCO数据集中GT目标与图像大小的IOU比例，中位数是 0.106，而ImageNet数据集对应比例的中位数是0.556。</li>
<li>经过分析后可以得知，COCO数据集中的目标尺寸变化范围非常大，这样在做迁移学习时，可能存在很多问题（即在ImageNet上预训练的模型用于coco数据集的检测）。</li>
</ul>
<p><img src="/pictures/SNIP1.png" alt="SNIP1"></p>
<!--noindex-->
<div class="post-button">
<a class="btn" href="/2020-09-02/%E5%B0%8F%E7%9B%AE%E6%A0%87%E6%A3%80%E6%B5%8B%E8%AE%BA%E6%96%87%E9%98%85%E8%AF%BB.html#more" rel="contents">
阅读全文 »
</a>
</div>
<!--/noindex-->
</div>
<footer class="post-footer">
<div class="post-eof"></div>
</footer>
</article>
<article itemscope itemtype="http://schema.org/Article" class="post-block" lang="zh-CN">
<link itemprop="mainEntityOfPage" href="http://lt4hyl.top/2020-09-02/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%885%EF%BC%89.html">
<span hidden itemprop="author" itemscope itemtype="http://schema.org/Person">
<meta itemprop="image" content="/images/avatar.gif">
<meta itemprop="name" content="刘涛">
<meta itemprop="description" content="为人民日益增长的美好生活需要而读书：机器学习、深度学习、c++、python、C#">
</span>
<span hidden itemprop="publisher" itemscope itemtype="http://schema.org/Organization">
<meta itemprop="name" content="Lancelot的小站">
</span>
<header class="post-header">
<h2 class="post-title" itemprop="name headline">
<a href="/2020-09-02/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%885%EF%BC%89.html" class="post-title-link" itemprop="url">深度学习笔记（5）</a>
</h2>
<div class="post-meta">
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar"></i>
</span>
<span class="post-meta-item-text">发表于</span>
<time title="创建时间：2020-09-02 15:01:52" itemprop="dateCreated datePublished" datetime="2020-09-02T15:01:52+08:00">2020-09-02</time>
</span>
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar-check"></i>
</span>
<span class="post-meta-item-text">更新于</span>
<time title="修改时间：2020-09-03 16:12:48" itemprop="dateModified" datetime="2020-09-03T16:12:48+08:00">2020-09-03</time>
</span>
<span id="/2020-09-02/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%885%EF%BC%89.html" class="post-meta-item leancloud_visitors" data-flag-title="深度学习笔记（5）" title="阅读次数">
<span class="post-meta-item-icon">
<i class="fa fa-eye"></i>
</span>
<span class="post-meta-item-text">阅读次数：</span>
<span class="leancloud-visitors-count"></span>
</span><br>
<span class="post-meta-item" title="本文字数">
<span class="post-meta-item-icon">
<i class="far fa-file-word"></i>
</span>
<span class="post-meta-item-text">本文字数：</span>
<span>7.7k</span>
</span>
<span class="post-meta-item" title="阅读时长">
<span class="post-meta-item-icon">
<i class="far fa-clock"></i>
</span>
<span class="post-meta-item-text">阅读时长 ≈</span>
<span>7 分钟</span>
</span>
</div>
</header>
<div class="post-body" itemprop="articleBody">
<h1 id="Sequence-Models"><a href="#Sequence-Models" class="headerlink" title="Sequence Models"></a>Sequence Models</h1><h2 id="RNN的引入"><a href="#RNN的引入" class="headerlink" title="RNN的引入"></a>RNN的引入</h2><p>在使用标准的神经网络（如：DNN）输入序列化数据（如：单词的 one-hot编码），可能存在的问题：</p>
<ul>
<li>输入和输出，对不同的实例会有不同长度，虽然可以用0填充到最大长度，但这个方法不好；</li>
<li>神经网络太简单，不会共享从不同文本位置学习到的特征。（输入前后的序列特征、上下文特征，没有学习到）。</li>
</ul>
<p><img src="/pictures/image-20200707111215551.png" alt="image-20200707111215551"></p>
<!--noindex-->
<div class="post-button">
<a class="btn" href="/2020-09-02/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%885%EF%BC%89.html#more" rel="contents">
阅读全文 »
</a>
</div>
<!--/noindex-->
</div>
<footer class="post-footer">
<div class="post-eof"></div>
</footer>
</article>
<article itemscope itemtype="http://schema.org/Article" class="post-block" lang="zh-CN">
<link itemprop="mainEntityOfPage" href="http://lt4hyl.top/2020-08-28/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%884%EF%BC%89.html">
<span hidden itemprop="author" itemscope itemtype="http://schema.org/Person">
<meta itemprop="image" content="/images/avatar.gif">
<meta itemprop="name" content="刘涛">
<meta itemprop="description" content="为人民日益增长的美好生活需要而读书：机器学习、深度学习、c++、python、C#">
</span>
<span hidden itemprop="publisher" itemscope itemtype="http://schema.org/Organization">
<meta itemprop="name" content="Lancelot的小站">
</span>
<header class="post-header">
<h2 class="post-title" itemprop="name headline">
<a href="/2020-08-28/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%884%EF%BC%89.html" class="post-title-link" itemprop="url">深度学习笔记（4）</a>
</h2>
<div class="post-meta">
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar"></i>
</span>
<span class="post-meta-item-text">发表于</span>
<time title="创建时间：2020-08-28 12:26:45" itemprop="dateCreated datePublished" datetime="2020-08-28T12:26:45+08:00">2020-08-28</time>
</span>
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar-check"></i>
</span>
<span class="post-meta-item-text">更新于</span>
<time title="修改时间：2020-09-03 09:42:41" itemprop="dateModified" datetime="2020-09-03T09:42:41+08:00">2020-09-03</time>
</span>
<span id="/2020-08-28/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%884%EF%BC%89.html" class="post-meta-item leancloud_visitors" data-flag-title="深度学习笔记（4）" title="阅读次数">
<span class="post-meta-item-icon">
<i class="fa fa-eye"></i>
</span>
<span class="post-meta-item-text">阅读次数：</span>
<span class="leancloud-visitors-count"></span>
</span><br>
<span class="post-meta-item" title="本文字数">
<span class="post-meta-item-icon">
<i class="far fa-file-word"></i>
</span>
<span class="post-meta-item-text">本文字数：</span>
<span>11k</span>
</span>
<span class="post-meta-item" title="阅读时长">
<span class="post-meta-item-icon">
<i class="far fa-clock"></i>
</span>
<span class="post-meta-item-text">阅读时长 ≈</span>
<span>10 分钟</span>
</span>
</div>
</header>
<div class="post-body" itemprop="articleBody">
<h1 id="Computer-Vision（计算机视觉）"><a href="#Computer-Vision（计算机视觉）" class="headerlink" title="Computer Vision（计算机视觉）"></a>Computer Vision（计算机视觉）</h1><h2 id="Convolutional-Neural-Networks"><a href="#Convolutional-Neural-Networks" class="headerlink" title="Convolutional Neural Networks"></a>Convolutional Neural Networks</h2><ul>
<li>随着感知向量机层数的加深，逐渐进入了深层感知向量机也可以叫做 <strong>深度神经网络（DNN）</strong>；</li>
<li>DNN在处理图片数据时会引发<strong>参数灾难</strong><ul>
<li>DNN的隐层之间都是采用<strong>全连接</strong>的形式，假设图片较大，比如：300x300x3（宽高均为300像素，红蓝绿三通道）；</li>
<li>此时假设DNN某一隐层的输入等于图片大小，输出也想等于图片的大小，则该层的参数数量就为：<code>300x300x3x300x300x3=72,900,000,000</code>，也就7百万的参数量，这将导致网络无法训练，也就是常说的<strong>参数灾难</strong>或者<strong>维度灾难</strong>。</li>
<li>因此需要引入其它的处理方式，降低参数量的同时还能有效的学习到图片中包含的信息，也就是<strong>CNN：卷积神经网络</strong>。</li>
</ul>
</li>
</ul>
<!--noindex-->
<div class="post-button">
<a class="btn" href="/2020-08-28/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%884%EF%BC%89.html#more" rel="contents">
阅读全文 »
</a>
</div>
<!--/noindex-->
</div>
<footer class="post-footer">
<div class="post-eof"></div>
</footer>
</article>
<article itemscope itemtype="http://schema.org/Article" class="post-block" lang="zh-CN">
<link itemprop="mainEntityOfPage" href="http://lt4hyl.top/2020-08-27/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%883%EF%BC%89.html">
<span hidden itemprop="author" itemscope itemtype="http://schema.org/Person">
<meta itemprop="image" content="/images/avatar.gif">
<meta itemprop="name" content="刘涛">
<meta itemprop="description" content="为人民日益增长的美好生活需要而读书：机器学习、深度学习、c++、python、C#">
</span>
<span hidden itemprop="publisher" itemscope itemtype="http://schema.org/Organization">
<meta itemprop="name" content="Lancelot的小站">
</span>
<header class="post-header">
<h2 class="post-title" itemprop="name headline">
<a href="/2020-08-27/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%883%EF%BC%89.html" class="post-title-link" itemprop="url">深度学习笔记（3）</a>
</h2>
<div class="post-meta">
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar"></i>
</span>
<span class="post-meta-item-text">发表于</span>
<time title="创建时间：2020-08-27 16:25:34" itemprop="dateCreated datePublished" datetime="2020-08-27T16:25:34+08:00">2020-08-27</time>
</span>
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar-check"></i>
</span>
<span class="post-meta-item-text">更新于</span>
<time title="修改时间：2020-09-02 15:05:43" itemprop="dateModified" datetime="2020-09-02T15:05:43+08:00">2020-09-02</time>
</span>
<span id="/2020-08-27/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%883%EF%BC%89.html" class="post-meta-item leancloud_visitors" data-flag-title="深度学习笔记（3）" title="阅读次数">
<span class="post-meta-item-icon">
<i class="fa fa-eye"></i>
</span>
<span class="post-meta-item-text">阅读次数：</span>
<span class="leancloud-visitors-count"></span>
</span><br>
<span class="post-meta-item" title="本文字数">
<span class="post-meta-item-icon">
<i class="far fa-file-word"></i>
</span>
<span class="post-meta-item-text">本文字数：</span>
<span>5.7k</span>
</span>
<span class="post-meta-item" title="阅读时长">
<span class="post-meta-item-icon">
<i class="far fa-clock"></i>
</span>
<span class="post-meta-item-text">阅读时长 ≈</span>
<span>5 分钟</span>
</span>
</div>
</header>
<div class="post-body" itemprop="articleBody">
<h1 id="机器学习策略-ML-Strategy"><a href="#机器学习策略-ML-Strategy" class="headerlink" title="机器学习策略(ML Strategy)"></a>机器学习策略(ML Strategy)</h1><ul>
<li>假设我们正在训练一个识别图像中是否包含猫咪的网络，为了提升网络的表现，我们可能会产生如下的想法：<ul>
<li>收集更多的数据；</li>
<li>收集更多不同种类猫的训练集；</li>
<li>使用梯度下降算法训练更长的时间</li>
<li>使用不同的优化算法，如：Adam代替原有的梯度下降算法。</li>
<li>尝试更大的/更小的网络；</li>
<li>使用 dropout/L2正则化 等；</li>
<li>尝试不同的网络架构；</li>
<li>……..等等</li>
</ul>
</li>
</ul>
<!--noindex-->
<div class="post-button">
<a class="btn" href="/2020-08-27/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%883%EF%BC%89.html#more" rel="contents">
阅读全文 »
</a>
</div>
<!--/noindex-->
</div>
<footer class="post-footer">
<div class="post-eof"></div>
</footer>
</article>
<article itemscope itemtype="http://schema.org/Article" class="post-block" lang="zh-CN">
<link itemprop="mainEntityOfPage" href="http://lt4hyl.top/2020-08-24/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%882%EF%BC%89.html">
<span hidden itemprop="author" itemscope itemtype="http://schema.org/Person">
<meta itemprop="image" content="/images/avatar.gif">
<meta itemprop="name" content="刘涛">
<meta itemprop="description" content="为人民日益增长的美好生活需要而读书：机器学习、深度学习、c++、python、C#">
</span>
<span hidden itemprop="publisher" itemscope itemtype="http://schema.org/Organization">
<meta itemprop="name" content="Lancelot的小站">
</span>
<header class="post-header">
<h2 class="post-title" itemprop="name headline">
<a href="/2020-08-24/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%882%EF%BC%89.html" class="post-title-link" itemprop="url">深度学习笔记（2）</a>
</h2>
<div class="post-meta">
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar"></i>
</span>
<span class="post-meta-item-text">发表于</span>
<time title="创建时间：2020-08-24 19:14:49" itemprop="dateCreated datePublished" datetime="2020-08-24T19:14:49+08:00">2020-08-24</time>
</span>
<span class="post-meta-item">
<span class="post-meta-item-icon">
<i class="far fa-calendar-check"></i>
</span>
<span class="post-meta-item-text">更新于</span>
<time title="修改时间：2020-09-02 15:05:22" itemprop="dateModified" datetime="2020-09-02T15:05:22+08:00">2020-09-02</time>
</span>
<span id="/2020-08-24/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AC%94%E8%AE%B0%EF%BC%882%EF%BC%89.html" class="post-meta-item leancloud_visitors" data-flag-title="深度学习笔记（2）" title="阅读次数">
<span class="post-meta-item-icon">
<i class="fa fa-eye"></i>
</span>
<span class="post-meta-item-text">阅读次数：</span>
<span class="leancloud-visitors-count"></span>
</span><br>
<span class="post-meta-item" title="本文字数">
<span class="post-meta-item-icon">
<i class="far fa-file-word"></i>
</span>
<span class="post-meta-item-text">本文字数：</span>
<span>3.7k</span>
</span>
<span class="post-meta-item" title="阅读时长">
<span class="post-meta-item-icon">
<i class="far fa-clock"></i>
</span>
<span class="post-meta-item-text">阅读时长 ≈</span>
<span>3 分钟</span>
</span>
</div>
</header>
<div class="post-body" itemprop="articleBody">
<h1 id="Improving-Deep-Neural-Networks-Hyperparameter-tuning-Regularization-and-Optimization"><a href="#Improving-Deep-Neural-Networks-Hyperparameter-tuning-Regularization-and-Optimization" class="headerlink" title="Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization"></a>Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization</h1><h2 id="Train-dev-test-sets"><a href="#Train-dev-test-sets" class="headerlink" title="Train/dev/test sets"></a>Train/dev/test sets</h2><ul>
<li>三者对总体数据量的划分问题：<ul>
<li>对于传统机器学习上，由于数据量较少，对数据的分割上可以将<code>Train/dev/test sets</code>分别分成：70/20/10%，也有 60/20/20%。</li>
<li>但对于现在的 <code>big data</code> 可以只用少部分数据作为 dev /test sets ，如：99/1/1%；就足以比较出来不同算法的性能好坏以及适用性。</li>
<li>其中对于测试集（test set）是在系统开发完成后，用于帮助我们评估最终系统的性能，因此测试集的大小只要足够能保证对系统整体性能评估的高置信度即可。</li>
</ul>
</li>
<li>三者数据的分布一致问题：<ul>
<li>首先最重要的是要：<strong>尽量保证 dev/test sets 中的数据分布是一致的</strong>；</li>
<li>使用 <code>dev sets</code>的目的也就是帮我们评估不同的想法，让我们能更好地从A或B中做出选择 。</li>
<li>使用 <code>test sets</code>的目的是为了保证对模型的无偏估计，也就是在系统开发完成后，用于帮助我们评估最终系统的性能，当然如果没有 test sets 的话 ，也是ok的。这将只在 train/dev sets 上进行训练和测试。</li>
</ul>