I am currently trying to recognise DVD covers in generic photos. My initial test involved using 100 DVD covers and 10 test cases of photos that contained them, and with some tweaking of the find_obj.cpp example in OpenCV I was able to get recognition working.
However now I need to do this on a much larger database, and I am aware that the FLANN method will not scale up well to meet this requirement. How do people here recommend I scale up my SURF recognition in an SQL database?
If you really want to scale your system to several orders of magnitude, nearest neighbors search (FLANN) will not be sufficient.
In such a case what you need is to build a visual vocabulary (a.k.a bag of words) by quantizing your descriptors, and create an inverted index.
I recommend you to refer to the Scalable Recognition with a Vocabulary Tree paper that is the reference publication for such a topic.
Related
I have a bunch of poor quality photos that I extracted from a pdf. Somebody I know has the good quality photo's somewhere on her computer(Mac), but it's my understanding that it will be difficult to find them.
I would like to
loop through each poor quality photo
perform a reverse image search using each poor quality photo as the query image and using this persons computer as the database to search for the higher quality images
and create a copy of each high quality image in one destination folder.
Example pseudocode
for each image in poorQualityImages:
search ./macComputer for a higherQualityImage of image
copy higherQualityImage to ./higherQualityImages
I need to perform this action once.
I am looking for a tool, github repo or library which can perform this functionality more so than a deep understanding of content based image retrieval.
There's a post on reddit where someone was trying to do something similar
imgdupes is a program which seems like it almost achieves this, but I do not want to delete the duplicates, I want to copy the highest quality duplicate to a destination folder
Update
Emailed my previous image processing prof and he sent me this
Off the top of my head, nothing out of the box.
No guaranteed solution here, but you can narrow the search space.
You’d need a little program that outputs the MSE or SSIM similarity
index between two images, and then write another program or shell
script that scans the hard drive and computes the MSE between each
image on the hard drive and each query image, then check the images
with the top X percent similarity score.
Something like that. Still not maybe guaranteed to find everything
you want. And if the low quality images are of different pixel
dimensions than the high quality images, you’d have to do some image
scaling to get the similarity index. If the poor quality images have
different aspect ratios, that’s even worse.
So I think it’s not hard but not trivial either. The degree of
difficulty is partly dependent on the nature of the corruption in the
low quality images.
UPDATE
Github project I wrote which achieves what I want
What you are looking for is called image hashing
. In this answer you will find a basic explanation of the concept, as well as a go-to github repo for plug-and-play application.
Basic concept of Hashing
From the repo page: "We have developed a new image hash based on the Marr wavelet that computes a perceptual hash based on edge information with particular emphasis on corners. It has been shown that the human visual system makes special use of certain retinal cells to distinguish corner-like stimuli. It is the belief that this corner information can be used to distinguish digital images that motivates this approach. Basically, the edge information attained from the wavelet is compressed into a fixed length hash of 72 bytes. Binary quantization allows for relatively fast hamming distance computation between hashes. The following scatter plot shows the results on our standard corpus of images. The first plot shows the distances between each image and its attacked counterpart (e.g. the intra distances). The second plot shows the inter distances between altogether different images. While the hash is not designed to handle rotated images, notice how slight rotations still generally fall within a threshold range and thus can usually be matched as identical. However, the real advantage of this hash is for use with our mvp tree indexing structure. Since it is more descriptive than the dct hash (being 72 bytes in length vs. 8 bytes for the dct hash), there are much fewer false matches retrieved for image queries.
"
Another blogpost for an in-depth read, with an application example.
Available Code and Usage
A github repo can be found here. There are obviously more to be found.
After importing the package you can use it to generate and compare hashes:
>>> from PIL import Image
>>> import imagehash
>>> hash = imagehash.average_hash(Image.open('test.png'))
>>> print(hash)
d879f8f89b1bbf
>>> otherhash = imagehash.average_hash(Image.open('other.bmp'))
>>> print(otherhash)
ffff3720200ffff
>>> print(hash == otherhash)
False
>>> print(hash - otherhash)
36
The demo script find_similar_images also on the mentioned github, illustrates how to find similar images in a directory.
Premise
I'll focus my answer on the image processing part, as I believe implementation details e.g. traversing a file system is not the core of your problem. Also, all that follows is just my humble opinion, I am sure that there are better ways to retrieve your image of which I am not aware. Anyway, I agree with what your prof said and I'll follow the same line of thought, so I'll share some ideas on possible similarity indexes you might use.
Answer
MSE and SSIM - This is a possible solution, as suggested by your prof. As I assume the low quality images also have a different resolution than the good ones, remember to downsample the good ones (and not upsample the bad ones).
Image subtraction (1-norm distance) - Subtract two images -> if they are equal you'll get a black image. If they are slightly different, the non-black pixels (or the sum of the pixel intensity) can be used as a similarity index. This is actually the 1-norm distance.
Histogram distance - You can refer to this paper: https://www.cse.huji.ac.il/~werman/Papers/ECCV2010.pdf. Comparing two images' histograms might be potentially robust for your task. Check out this question too: Comparing two histograms
Embedding learning - As I see you included tensorflow, keras or pytorch as tags, let's consider deep learning. This paper came to my
mind: https://arxiv.org/pdf/1503.03832.pdf The idea is to learn a
mapping from the image space to a Euclidian space - i.e. compute an
embedding of the image. In the embedding hyperspace, images are
points. This paper learns an embedding function by minimizing the
triplet loss. The triplet loss is meant to maximize the distance
between images of different classes and minimize the distance between
images of the same class. You could train the same model on a Dataset
like ImageNet. You could augment the dataset with by lowering the
quality of the images, in order to make the model "invariant" to
difference in image quality (e.g. down-sampling followed by
up-sampling, image compression, adding noise, etc.). Once you can
compute embedding, you could compute the Euclidian distance (as a
substitute of the MSE). This might work better than using MSE/SSIM as a similarity indexes. Repo of FaceNet: https://github.com/timesler/facenet-pytorch. Another general purpose approach (not related to faces) which might help you: https://github.com/zegami/image-similarity-clustering.
Siamese networks for predicting similarity score - I am referring to this paper on face verification: http://bmvc2018.org/contents/papers/0410.pdf. The siamese network takes two images as input and outputs a value in the [0, 1]. We can interpret the output as the probability that the two images belong to the same class. You can train a model of this kind to predict 1 for image pairs of the following kind: (good quality image, artificially degraded image). To degrade the image, again, you can combine e.g. down-sampling followed by
up-sampling, image compression, adding noise, etc. Let the model predict 0 for image pairs of different classes (e.g. different images). The output of the network can e used as a similarity index.
Remark 1
These different approaches can also be combined. They all provide you with similarity indexes, so you can very easily average the outcomes.
Remark 2
If you only need to do it once, the effort you need to put in implementing and training deep models might be not justified. I would not suggest it. Still, you can consider it if you can't find any other solution and that Mac is REALLY FULL of images and a manual search is not possible.
If you look at the documentation of imgdupes you will see there is the following option:
--dry-run
dry run (do not delete any files)
So if you run imgdupes with --dry-run you will get a listing of all the duplicate images but it will not actually delete anything. You should be able to process that output to move the images around as you need.
Try similar image finder I have developed to address this problem.
There is an explanation and the algorithm there, so you can implement your own version if needed.
In 2007, when I was young and foolish and before I knew about Open Street Map, I started an urban historical map project. I was working in Illustrator, it was going to be an interactive Flash piece, and my process was to draw the maps first, with the thought that I'd label some, but not all, of the street later on.
As we know Flash was began to die about 2010 and I put the project away for a number of years. I picked it up again a couple years ago and continued my earlier practice of just drawing streets and water features, this time with the intention of making it a conventional web map. Now I'm pretty close to finishing the drawing of a five-layer (1871, 1903, 1932, 1952 and 2016) historical map of a medium-sized city, though it still lacks labels.
My problem now is how to add large numbers of labels, many of them duplicates. There could be as many as 10,000 for all five layers, though as a practical matter I may have to settle for a smallish fraction of that number. Based on web searches I gather my workflow is unusual and that mine is therefore an unusual problem.
I've exported my maps and brought them into QGIS and played with the software a little. The process of adding labels to objects doesn't seem terribly efficient or user-friendly, but that's probably due to my unfamiliarity with the program.
So my question is this: Are there any tricks to speed up the painful process of adding large numbers of duplicate labels in either QGIS or ArcGIS? Since so many of the streets exist in all five layers, functionality like the ability to select multiple objects in different layers and edit their attributes simultaneously in the Attribute Table would be a godsend. (Doesn't seem possible.) So would the ability to copy the attributes from one object and paste them onto other objects. Or the ability to do either of these things in Illustrator via a plugin and then export the data along with the shapes to a GIS program.
Thanks for your help!
If I understand the issue correctly I think are several different solutions. When you say that you
Typically for a spatial layer in ArcGIS or QGIS you define how to label all features in a layer once by defining a label scheme to use across all features, 1 or 1 million. This assumes that each feature in the layer has one or more attributes in the associated table for the layer.
How are you converting the Illustrator vectors to a spatial layer? DXF?
You will likely have better/faster responses to this question by posting it to the GIS Stack exchange. https://gis.stackexchange.com/
https://learn.microsoft.com/fr-fr/azure/cognitive-services/computer-vision/concept-object-detection
https://cloud.google.com/vision/docs/object-localizer
I would want to know how many and which objects are recognizable using theses APIs and I can't find a mention of that fact.
I found that google API use https://developers.google.com/knowledge-graph/ which is based on schema.org types but I don't really understand well what it's all about.
I'm sorry but as far as I know, there is no fixed list of classes that Azure Computer Vision is able to detect.
By the way, even if there was one, this list is evolving on a regular basis (but no schedule is announced).
In any case, there are limitations (see doc here):
It's important to note the limitations of object detection so you can
avoid or mitigate the effects of false negatives (missed objects) and
limited detail.
Objects are generally not detected if they're small (less than 5% of
the image).
Objects are generally not detected if they're arranged
closely together (a stack of plates, for example).
Objects are not differentiated by brand or product names (different
types of sodas on a store shelf, for example). However, you can get
brand information from an image by using the Brand detection feature.
If you want to detect specific objects, I would highly suggest using Custom Vision (doc / overview here), not Computer Vision, where you can train your model with your own images to match what you are trying to detect
I am making a little personal project.
Ideally I would like to be able to make programmatically a google search and have the count of results. (My goal is to compare the results count between a lot (100000+) of different phrases).
Is there a free way to make a web search and compare the popularity of different texts, by using Google Bing or whatever (the source is not really important).
I tried Google but seems that freely I can do only 10 requests per day.
Bing is more permissive (5000 free requests per month).
Is there other tools or way to have a count of number of results for a particular sentence freely ?
Thanks in advance.
There are several things you're going to need if you're seeking to create a simple search engine.
First of all you should read and understand where the field of information retrieval started with G. Salton's paper or at least read the wiki page on the vector space model. It will require you learning at least some undergraduate linear algebra. I suggest Gilbert Strang's MIT video lectures for this.
You can then move to the Brin/Page Pagerank paper which outlays the original concept behind the hyperlink matrix and quickly calculating eigenvectors for ranking or read the wiki page.
You may also be interested in looking at the code for Apache Lucene
To get into contemporary search algorithm techniques you need calculus and regression analysis to learn machine learning and deep learning as the current google search has moved away from Pagerank and utilizes these. This is partially due to how link farming enabled people to artificially engineer search results and the huge amount of meta data that modern browsers and web servers allow to be collected.
EDIT:
For the webcrawler only portion I'd recommend WebSPHINX. I used this in my senior research in college in conjunction with Lucene.
We currently have a dynamically updated network graph with around 1,500 nodes and 2,000 edges. It's ever-growing. Our current layout engine uses Prefuse - the force directed layout in particular - and it takes about 10 minutes with a hefty server to get a nice, stable layout.
I've looked a little GraphViz's sfpd algorithm, but haven't tested it yet...
Are there faster alternatives I should look at?
I don't care about the visual appearance of the nodes and edges - we process that separately - just putting x, y on the nodes.
We do need to be able to tinker with the layout properties for specific parts of the graph, for instance, applying special tighter or looser springs for certain nodes.
Thanks in advance, and please comment if you need more specific information to answer!
EDIT: I'm particularly looking for speed comparisons between the layout engine options. Benchmarks, specific examples, or just personal experience would suffice!
I wrote a JavaScript-based graph drawing library VivaGraph.js.
It calculates layout and renders graph with 2K+ vertices, 8.5K edges in ~10-15 seconds. If you don't need rendering part it should be even faster.
Here is a video demonstrating it in action: WebGL Graph Rendering With VivaGraphJS.
Online demo is available here. WebGL is required to view the demo but is not needed to calculate graphs layouts. The library also works under node.js, thus could be used as a service.
Example of API usage (layout only):
var graph = Viva.Graph.graph(),
layout = Viva.Graph.Layout.forceDirected(graph);
graph.addLink(1, 2);
layout.run(50); // runs 50 iterations of graph layout
// print results:
graph.forEachNode(function(node) { console.log(node.position); })
Hope this helps :)
I would have a look at OGDF, specifically http://www.ogdf.net/doku.php/tech:howto:frcl
I have not used OGDF, but I do know that Fast Multipole Multilevel is a good performant algorithm and when you're dealing with the types of runtimes involved with force directed layout with the number of nodes you want, that matters a lot.
Why, among other reasons, that algorithm is awesome: Fast Multipole method. The fast multipole method is a matrix multiplication approximation which reduces the O() runtime of matrix multiplication for approximation to a small degree. Ideally, you'd have code from something like this: http://mgarland.org/files/papers/layoutgpu.pdf but I can't find it anywhere; maybe a CUDA solution isn't up your alley anyways.
Good luck.
The Gephi Toolkit might be what you need: some layouts are very fast yet with a good quality: http://gephi.org/toolkit/
30 secondes to 2 minutes are enough to layout such a graph, depending on your machine.
You can use the ForAtlas layout, or the Yifan Hu Multilevel layout.
For very large graphs (+50K nodes and 500K links), the OpenOrd layout wil
In a commercial scenario, you might also want to look at the family of yFiles graph layout and visualization libraries.
Even the JavaScript version of it can perform layouts for thousands of nodes and edges using different arrangement styles. The "organic" layout style is an implementation of a force directed layout algorithm similar in nature to the one used in Neo4j's browser application. But there are a lot more layout algorithms available that can give better visualizations for certain types of graph structures and diagrams. Depending on the settings and structure of the problem, some of the algorithms take only seconds, while more complex implementations can also bring your JavaScript engine to its knees. The Java and .net based variants still perform quite a bit better, as of today, but the JavaScript engines are catching up.
You can play with these algorithms and settings in this online demo.
Disclaimer: I work for yWorks, which is the maker of these libraries, but I do not represent my employer on SO.
I would take a look at http://neo4j.org/ its open source which is beneficial in your case so you can customize it to your needs. The github account can be found here.