b******g
发帖数: 3616 | 1 你发的第一张看上去像HDR.
Photomatix我用了很多年了,从3.0版用到现在的5.0版.这个软件还是比较简单的,
tutorial网上到处都是,包括这个公司的官方网站都有.但是真正用得好的人还是不多.
很多做出来还是太假.
至于ps的blend,你有没有试过mask的时候用image->apply image? |
|
o**s
发帖数: 3412 | 2 Thanks for the feedback people.
I am gonna give photoshop->merge to HDR pro and manual blending another try,
with more patience and look into Photomatix.
No have not tried image->apply image. |
|
b******g
发帖数: 3616 | 3 你可以试下对mask进行image->apply image
这是一个挺有意思的技术,就是把原来的图像加到mask上,根据原始图像的亮度来确定
mask上的灰度.这样一来,效果就是根据照片本身把暗的地方自然的mask了,如果你对欠
暴那
张采用这个mask,那么他的高光部分(比如天空)就被保留下来了.同样,你把mask
invert一下,对过暴那张使用,就保留了阴影部分.这个技术提供了一个很好的mask的
starting point.当然这个和HDR是没有关系的.纯粹是手工blend.
我的建议是,如果你走HDR的话,就用photomatix吧,是最专业的HDR软件,几乎没有之一.
try, |
|
|
|
y*b
发帖数: 3190 | 6 Digicame-info leaked image and specs of the EF 11-24mm f/4L USM lens. The
lens will be announced next week.
Specs:
– Ultra-wide-angle zoom lens for full-frame
– Dust and water sealed
– SWC coating
– ASC coating
– Fluorine coating
– One large diameter grinding aspheric lens, one large-diameter glass
molded aspherical lens, two glass molded aspherical lens, one super UD
lenses, one lens UD
http://www.canonwatch.com/cw5-canon-ef-11-24mm-f4l-usm-image-an |
|
|
y*b
发帖数: 3190 | 8 zz from CW
"Is this Canon patent application describing a mirrorless system camera with
an EVF and IBIS? We tend to believe it does.
Canon patent application US20180041705 refers clearly to image stabilisation
“in camera”. Moreover, as one reader noted, it looks like it will use a
current Canon mount, since it doesn’t have the short flange a mirrorless
must would have. Furthermore, there is no mirror in the images. It appears
that an electronic viewfinder is placed where usually the pentaprism i... 阅读全帖 |
|
|
x*****g
发帖数: 2 | 10 Dear friends,
The majoy subject of my sister is on Medical imaging
[Ying(3)Xiang(4)] for five years. And now she want to
apply
to study in Universities of USA, however we don't know the
situation of her major. PLEASE give us some suggestions or
advice on it if you have the time! Whether medical imaging
is easy to apply in USA and if you know, which universities
she should check? If she want to change her major, which
subjects are better to find the solution and
easy to obtain the sch |
|
c*******h
发帖数: 1096 | 11 there are many image search engines, all of which rely on
text and metadata, including this advertised method by google.
without text, pure imaged based search is impossible. believe me.
---from an artificial-intellegence-pesimmist. |
|
ET
发帖数: 10701 | 12 that's what google's article talked about. they provide a image search
approach, implement it, that's what I understand.
if the image search has to be based on text, metadata , label, what's points
that the google, and other companies make those efforts? |
|
c*******h
发帖数: 1096 | 13 the possibility: without an initial step of retrieving a set of images
based on texts and tags, the random surfer model and the image feature
method proposed by the paper is impractical---that's what the paper is
about.
my point is that while impressed by the model and method advertised by
the paper, you should know the key requirement for it to work---texts
and tags. however, this latter part---the restriction---tends to be
overlooked by people.
is
to |
|
I**A
发帖数: 2345 | 14 【 以下文字转载自 Programming 讨论区 】
发信人: IAYA (你是我的天使), 信区: Programming
标 题: 来measure web page (including image) loading time的软件
发信站: BBS 未名空间站 (Thu Dec 2 22:31:01 2010, 美东)
我用的YSLOW感觉检测的是server的response time。
我想检测的是从URL按回车到charts image在client browser显示出来之间的这段时间
大家有没有人用过类似的？
多谢！ |
|
w*******u
发帖数: 131 | 15 SUMMARY OF POSITION:
Image Mastering is responsible for the color mastering of Pixar projects -
including features, shorts, promotional pieces, and DVD materials – and
creation of the image-based distribution masters for domestic and
international release. The color mastering begins on an in-house digital
color grading system and continues through the delivery formats, which
include film, digital cinema, and video (standard and high definition). We
are also responsible for the color science requ |
|
s*****a
发帖数: 152 | 16 【 以下文字转载自 EE 讨论区,原文如下 】
发信人: sighaaa (sighaaa), 信区: EE
标 题: A Question about image registration
发信站: Unknown Space - 未名空间 (Sun Aug 1 19:47:11 2004) WWW-POST
Here I have a simple question about image registration. Is Euler 3D Transform
a non-linear transform? Thanks. |
|
h***n
发帖数: 276 | 17 【 以下文字转载自 EE 讨论区 】
【 原文由 jinghanna 所发表 】
I have a sequence of grayscale images, and want to write them into one image
file, for example *.tiff file. How can I do this?
I know maybe I can use the matlab function imwrite, but when I use this func
tion, I meet some problem.
I use the following command,
imwrite(X,map,filename,fmt)
How can I determine the map term above?
Thanks a lot for your kind help! |
|
x*x
发帖数: 7 | 18 CV=image processing+3D modeling+camera vision+video processing etc.
有时候image processing特指信号处理的一种，如果在EE看来。 |
|
a*******a
发帖数: 8 | 19 SPIE,
IEEE Medical Imaging conference,
if you are doing emission imaging
Society of Nuclear Medicine annual meeting |
|
P********e
发帖数: 2610 | 20 Anyone has any advice?
Basically, I'll need to generate barcode( which is easy)
And, take pictures(images) and to decode the barcode to numbers(or letters)
So, my question is how to decode the barcode by using images? |
|
p*****n
发帖数: 371 | 21 正在准备选择PHD的研究方向。因为导师就是搞IMAGE/VISION这块的。
他建议我PHD论文搞Digital Image Forensics这个方向。
我感觉这个方向搞的人好像不是很多，另外应用领域似乎也不是很宽，不知道前景到底
如何？听导师的意思好像将来出去主要是去金融和政府机构。
我个人还是比较喜欢像MSFT或者GOOGLE这样的大型软件开发公司。
不知道这里有没有大牛给些建议和意见？
谢谢了！ |
|
S*******w
发帖数: 24236 | 22 这些公司编程算法牛就行了
跟搞啥没关系
正在准备选择PHD的研究方向。因为导师就是搞IMAGE/VISION这块的。
他建议我PHD论文搞Digital Image Forensics这个方向。
我感觉这个方向搞的人好像不是很多，另外应用领域似乎也不是很宽，不知道前景到底
如何？听导师的意思好像将来出去主要是去金融和政府机构。
我个人还是比较喜欢像MSFT或者GOOGLE这样的大型软件开发公司。
不知道这里有没有大牛给些建议和意见？
谢谢了！ |
|
l*******r
发帖数: 511 | 23 【 以下文字转载自 JobHunting 讨论区 】
发信人: littlemar (小火星), 信区: JobHunting
标 题: 有没有人说说image processing之类的还有什么比较新的热门点
发信站: BBS 未名空间站 (Tue May 4 22:16:28 2010, 美东)
发信人: littlemar (小火星), 信区: CS
标 题: 有没有人说说image processing之类的还有什么比较新的热门点
发信站:
BBS 未名空间站 (Tue May 4 22:12:02 2010, 美东)
或者有待解决的。虚心求教。谢谢啦 |
|
r***h
发帖数: 70 | 24 小弟准备1年后毕业，对工业界情况不太很了解。请
问各位朋友哪些公司在做Medical Imaging,
Computer Vision的东西。我该去申请哪些职位？
面试前应该准备哪些东西？
我的情况大致如下。国内还不错的学校cs本科，美
国50-100名master,现ph.d在读。RA几年做
Medical Imaging, Pattern Recognition,
Computer Vision，4篇一作文章，除了1
篇Transaction的期刊外，都是很一般的会
议，估计毕业时可能会有6篇一作。在美国一家小公
司混过1年的software engineer，公司太小，也
没长什么见识。没其他什么工作经验。编程上只熟悉
C++和Matlab。
当时选这个方向是比较喜欢做些有idea的东西。
有人说这个方向好找工作，有人说不好找。当时也没
想太多就上手做了，觉得还蛮有意思。所以倾向于做
些R&D的Coding，纯粹的Coding的工作也愿意接
受。
请教各位有经验的朋友，有哪些公司在做这个方向，
我的背景该去申请什么样的职位？ |
|
r***h
发帖数: 70 | 25 希望您老给些建议，不要竟说风凉话。这3个是CV方
向的，做过的都知道，跟做medical imaging没多
大关系，比如Trans. Medical Imaging要比这3
个tough。
谢谢！
center的话 没有几篇拿的出手的paper( |
|
d******e
发帖数: 7844 | 26 Medical Imaging本来也不是啥好找工作的行当。
估计LZ也没有什么实习经验，去一去GE,SIMENS之类的地方就知道Medical Imaging是个
多么小的圈子了，就那几个大公司玩得起。
oriented
for |
|
r***h
发帖数: 70 | 27 Thank you so much for your suggestions.
future) 的大作中提到: 】
during your years in medical imaging,
you
skills in signal processing, image
intelligence, statistics, high-
many high-paying software development
employment opportunities are outside
the
the size of this industry is too small,
regulated. Only big companies can have
the
approval to develop medical software. |
|
m******0
发帖数: 90 | 28 请问各位大牛，大学里生物图像处理方向（image analysis, image processing）怎么
样？能够描述一下工作中具体做什么，一般用什么语言和技能，需要什么基础？
我只是知道需要matlab和machine learning, 但是不知道具体的实现过程。
做这个方向以后去工业界好找工作吗？
请前辈们不吝赐教！
多谢！！！！！！！ |
|
A**********e
发帖数: 3102 | 29 say i have an image file 'test.jpg' on my local linux machine, and want to
use PL/SQL and sqlplus to send it to a database on a server.
i know how to do it with C/C#, and how to do it with SQL if the image is on
the server. But how to do it with SQL on local machine?
THANKS~~~ :)) |
|
s***r
发帖数: 2604 | 30 so, basically without image structure info I can not fetch the image out?
even I already have theaddress, width and height of it?
can u give some marshall related link about this. I tried some marshall.copy
(), but not work. I am worry about something wrong in my code. |
|
k***e
发帖数: 7933 | 31 如果把装系统的硬盘换了，换之前做的系统image
（true image做的）还能在新的硬盘里面还原吗？ |
|
w*********t
发帖数: 928 | 32 要是file-based,坏了一点。别的还在。
要是image-based坏了一点，是不是就全完了?
另外，file-based, 外置硬盘，插到别的机器上，就能读出来。image-based是不是看软
件呀？ |
|
r***y
发帖数: 4379 | 33 系统分区和c一起克隆?
ghost 做不到啊. 只能每个分区搞一个image, 这情况就是分别搞两个images .
或者整盘clone, 可笔记本太难搞整盘clone. 用external enclosure 老是掉电. |
|
y**b
发帖数: 10166 | 34 rebate这种情况都是故意找借口拒绝的，不然那些rebate公司赚啥，
你痛骂几句马上给你通过。
但这个软件这几年实在烂到家了，早恨得牙痒痒，0daydown有最新
版破解，我都不愿意用。
今天测试了macrium reflect free, 完全可以取代true image，
立即卸载几台机器上的true image。 |
|
p********f
发帖数: 787 | 35 【 以下文字转载自 Java 讨论区 】
发信人: puffendorf (福宝), 信区: Java
标 题: Image processing library
发信站: BBS 未名空间站 (Tue Mar 27 11:08:27 2007)
Need to do some simple image processing (encoding, resizing, cropping, etc.)
. What's the best library out there? IamgeMagick? GD? Thanks. |
|
d****n
发帖数: 5 | 36 没接触过Java Image。
要作一个Servlet程序，需要动态的生成一些PNG图像，以及放大缩小和颜色操作。
大概看了一下Java Advanced Imaging API还有一些AWT API。不知道Java的performance
怎么样。有没有比较好的开放原码工具？请高人指点。多谢 |
|
d****n
发帖数: 5 | 37 Thanks.
If Java's performance is really bad, I will have to use JNI to connect to
other tools. One example is to use JMagick - ImageMagick. But I really want a
pure Java solution if possible.
Java Image APIs confused me a lot. We have Java2D, JAI (Java Advanced Imaging
API), new ImageIO API and some old APIs in AWT package. Hope some experts can
give me hints so I can shortcut to the right direction.
择
API。不知道Java的performance |
|
g*****g
发帖数: 34805 | 38 JAI is not a part of J2SE. Like other packages such as Java3D, they
provide a Sun implementation of some multimedia processing utilities.
If you care about platform-neutralism, you have to stick with J2SE,
which provides very limited support in image processing. ImageIO API
is new in J2SE1.4, you will have compatible problem if some of your clients
are still using 1.3 or lower. However, it looks like you are gonna process
image only on server end. So that's not a big deal. Just go ahead and
use |
|
o*******r
发帖数: 3 | 39 JimiRasterImage jrf
= Jimi.createRasterImage( );
if (fileType == 1)
{
Jimi.putImage("image/jpg", jrf,
new FileOutputStream(fileName));
}
else if (fileType == 3){
Jimi.putImage("image/png", jrf,
new FileOutputStream(fileName));
} |
|
d*r
发帖数: 238 | 40 I tried Image.getScaledInstance() and use Image.SCALE_SMOOTH as parameter. But
it is too slow.
些 |
|
F****n
发帖数: 3271 | 41 首先, 这取绝于你的目的到底是什么, 如果你就是想ZOOM IN/OUT而不对IMAGE进行修改,
那么800/600放到1024/768肯定有锯齿, 没有剧齿才是不正确的, 用AffineTransform就行
了,
如果你想修改IMAGE, 那么必然要CREATE一个SCALED的INSTANCE, 肯定会有一些计算, 会
花一些处理时间, 但也是正常的, JAVA一般在这些地方不会比C慢.
But
有 |
|
s******n
发帖数: 34 | 42 I can load .jpg, .gif files using ImageIO.read(),
and another class for loading .bmp image, which is found from somewhere.
how can I load a .tif image?
Many Thanks!!! |
|
g*****g
发帖数: 34805 | 43 You don't, but on the other hand, if there's no security issue,
I'd seriously consider caching images in appserver. And only goto
DB if image cannot be found in cache. |
|
h*******c
发帖数: 248 | 44 My image have vmware-tools installed, also hardware drivers are configured
for vmware.I didn't even install the generic video driver for xorg, but just
the vm driver. That is the reason X cannot work with your system.
try the following
yum install xorg-x11-drv-vesa
then you should have a generic X driver.
To build an image from minimal install of F10, then add X and fluxbox could
be cleaner for virtualbox.
Haven't try virtualbox yet. Does it have good communication with the host
system? I mean c |
|
w*s
发帖数: 7227 | 45 in windows, you can build the image, and its pdb file (which is used for
debugging.)
so you ship a small size of image to customer.
if it crashes, you have the pdb file in house for debugging.
how about linux ? |
|
c***d
发帖数: 996 | 46 ☆─────────────────────────────────────☆
piratenoob (pirate) 于 (Mon Feb 5 16:57:39 2007) 提到:
LD要我帮忙写一个应用程序（生物实验用的）但是我摸不到头绪，不知道从哪开始，
特来求教具体问题如下：
LD现在要观察小老鼠的行为，需要纪录小老鼠对某个区域A,B,C,D..的访问时间(7:
30,8:40...)。ABCD四个地方有四个camera 24小时监控。除了老鼠访问期间，其他的图
像数据都是junk.我想写一个程序能够对image做比较来判定是否有小老鼠来访问。这样
就可以剪裁video只记录我需要的部分。能给LD省很多时间干别的。不知道是不是
doable.
现在的困难是：
1）什么软件可以把整段video转化成一连串的image
2)有没有相关的paper对这种comparison有好的算法？
3）google的关键字……（Iam totally a noob at this area)
多谢大侠帮忙
☆───────────────────────────────── |
|
w***g
发帖数: 5958 | 47 Google用什么方法我不知道。但是分析搜索结果我猜用了下面几种技术：
1. reverse image search，搜同一个图的不同版本. 用SIFT, SURF等local feature。
我的那个软件就是做这个的。
2. search by visual similarity. 可以简单地想像成按color histogram搜。Google
可能有比color histogram牛得多的特征。
3. search by surrounding text。
4. 可能还有别的零散的技术，比如人脸识别对象识别等等，从图片中抽取关键字。考
虑到Google财大气粗，甚至有能力对人进行建模，检测到人以后可以对前景/背景分别
提取特征。最近应该也用上了neural network吧。但neural network给出的标签应该是
非常high-level的，对搜索帮助不会特别大。
最后把所有的结果rank一下，出来大概就是这个结果。Google确实是牛。我见过一眼负
责image search的那个老头，一脸苦B的样子，可能压力也非常大。
结果 |
|
h*********u
发帖数: 30 | 48 Dear Friends:
I'm in serious trouble right now, and hope someone can help me. The technician
in my department tried to re-install everything on my laptop, and he told he
backed up my files on Symantec image. But later I found out that many of those
backup files are corrupted, i.e. with "0 byte" and a new date (8/8/05). That
technician said that Symantec provides no means for reparing those damaged
images.
What can I do? Please help me. There are so many important files on that
laptop.
Truly, and |
|
s*******k
发帖数: 71 | 49
Convert PS to EPS (encapsulated postscript) and Powerpoint would be able to
recognize it, or use WMF format. In principle, any vector-format image files
would be better than bitmap image files.
SH |
|
