Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision Both sides next revision
imagealgorithms:functionalities_new [2017/04/20 09:45]
pdufour [1. Introduction]
imagealgorithms:functionalities_new [2017/04/12 12:17] (current)
pdufour [2.1.1. Overview]
Line 3: Line 3:
 ====== - Introduction ====== ====== - Introduction ======
  
-JASTEC France delivers image recognition technologies via his product LTU Engine. The solution is available via licensed software or via the hosted platform: LTU Engine ​OnPremise/OnDemand+JASTEC France delivers image recognition technologies via his product LTU Engine. The solution is available via licensed software or via the hosted platform: LTU Engine ​ON Premise/ON Demand
 \\  \\ 
 \\  \\ 
-LTU Engine includes two distinct ​images ​processing ​technologies:+LTU Engine includes two images ​recognition ​technologies ​- Visual Search and Image Processing. 
 +\\ 
 +The Image Processing offers a Fine Images Comparison solution.
 \\  \\ 
-  * **[[functionalities#​visual_search_solutions|The Visual Search]]** that is divided into two recognition solutions – the image matching and the visual similarity search. +The Visual Search is divided into two standard ​recognition solutions –  the image matching and  the visual similarity search. 
-  * **[[functionalities#​fine_comparison_of_images|The Image Processing]]** that offers a Fine Images Comparison solution. +\\
 \\ \\
 These principal functionalities are packaged in LTU Engine, which provides the components necessary for creating and managing visual search applications,​ including JSON API and a comprehensive Administrative Interface. ​ These principal functionalities are packaged in LTU Engine, which provides the components necessary for creating and managing visual search applications,​ including JSON API and a comprehensive Administrative Interface. ​
Line 19: Line 20:
  
 In this document, you are given details about: ​ In this document, you are given details about: ​
-  ***[[functionalities#​image_matching_technology|Matching search]]**  +  ***image matching technology**  
-  ***[[functionalities#​similarity_search|Similarity ​search]]** +  ***visual similarity ​search** 
-  ***[[functionalities#​color_search|Color ​search]]** +  ***color search** 
-  ***[[functionalities#​fine_comparison_of_images|Fine Images Comparison]]** +  ***fine images comparison** 
  
 ====== - Visual Search Solutions ====== ====== - Visual Search Solutions ======
-===== - Overview ===== + 
-The visual search ​solutions ​allow to find, from a query image, identical or similar visuals in images databases. The search is based on object recognition,​ shape or color  and depends on upon the content of an image, rather than on textual information.+The visual search ​technologies ​allow to find, from a query image, identical or similar visuals in images databases. The search is based on object recognition,​ shape or color  and is based upon the content of an image, rather than based on textual information.
  
 \\ \\
-{{ :imagealgorithms:​ltu-engine-user-guide-001.jpg?​nolink&​ |}}+{{ :engine:6.0:​ltu-engine-user-guide-001.jpg?​nolink&​ |}}
 /* {{ :​imagealgorithms:​algo-0092.jpeg |Visual similarity search}} */ /* {{ :​imagealgorithms:​algo-0092.jpeg |Visual similarity search}} */
 \\ \\
-The visual search ​is composed by two key steps:+Two key technologies of visual search ​are image description and image search:
  
-  ***Indexation**:​ The first step toward making an image searchable is to create a descriptor of the image content. LTU Engine computes a visual signature for every image that describe its visual content in terms of color, shape, texture and many higher order visual features and stores it in a reference database. These descriptors are also called image DNAs.+  ***Image description (Indexation)**: The first step toward making an image searchable is to create a descriptor of the image content. LTU Engine computes a visual signature for every image that describe its visual content in terms of color, shape, texture and many higher order visual features and stores it in a reference database. These descriptors are also called image DNAs. Indexing time thus refers to the time it takes to complete this computation (usually given in time per image).
  
-  ​***Retrivial**: ​special comparison technology by which an image signature can be compared at extremely high speed with other image signatures ​from a database up to millions ​of images.+   * **Image search (Retrivial)**: Image search is the comparison of one image signature with another, or with millions of other signatures. LTU's proprietary architecture includes a special comparison technology by which an image signature can be compared at extremely high speed with other image signatures. Retrieval time thus refers ​to the time this computation takes (usually given in time per comparison or numbers ​of images ​that can be searched per second).| 
 +\\ 
 +The core data for LTU Engine is the signature not the image itself.
 {{ :​imagealgorithms:​visual_search_process.png?​800 |}} {{ :​imagealgorithms:​visual_search_process.png?​800 |}}
- +Each query return ​a references list and for each reference imageits score.
-Each search returns ​a references list, their distance (or score), optional keywords as well as additional algorithm details. +
- +
-===== - Visual distance ===== +
-The distance is an indicator for the relevance of the retrieved images: the closer the value to 0.0, the closer the retrieved image shares the same visual content as the query image.  +
- +
-{{gdraw>​16dDrwjI5aE7M8nSlikLvg0xEd_fuh6z0OD3OAPGU99k width=791 title="​test title" center}} +
-/* https://​docs.google.com/​drawings/​d/​16dDrwjI5aE7M8nSlikLvg0xEd_fuh6z0OD3OAPGU99k/​edit */+
 \\ \\
 + 
  
-The visual distance is normalized such that a visual distance: 
-  * equal to 0 is a clone 
-  * below 1.0 indicates a match 
-  * between 1.0 and 1.8 reveals a similarity 
- 
-\\ 
-{{ :​imagealgorithms:​algo-001.jpg?​nolink&​1000 |}} 
 ===== - Image Matching Technology ===== ===== - Image Matching Technology =====
 ==== - Overview ==== ==== - Overview ====
  
-The image matching technology is used to find, in database(s),​ images that: +The image matching technology is used to find, in database(s),​ images that are
-      * Look exactly ​the same (e.g. for deduplication) +      * Exactly ​the same (e.g. for deduplication) 
-      * Have been edited ​in any way (e.g. for tracking on copyright images) +      * Edited ​in any way (e.g. for tracking on copyright images) 
-      * Are photos ​taken of the same visual content (e.g. for print to mobile applications)+      * Photos ​taken of the same visual content (e.g. for print to mobile applications)
 \\ \\
  
 {{ :​imagealgorithms:​matching1-ex-fond-gris.png?​nolink&​1000 |}} {{ :​imagealgorithms:​matching1-ex-fond-gris.png?​nolink&​1000 |}}
 +
 +\\
 +
 +To identify matches between a query picture and visual references, the technology compares images signature.
 +An image signature (also called image DNA) is computed from distinct local image features. Each signature is unique.
 +
 ==== - Image Transformations ==== ==== - Image Transformations ====
  
-LTU Engine’s image matching ​technology ​is robust against several types of image transformations,​ detecting not only the exact same image, but also modified versions of the original image and object matches (photographs of same object).+LTU engine’s image matching ​functionality ​is robust against several types of image transformations,​ detecting not only the exact same image, but also modified versions of the original image and object matches (photographs of same object).
 \\  \\ 
 \\ \\
Line 74: Line 70:
 \\ \\
 \\ \\
-This part illustrates the types of image transformations that LTU Engine ​can handle in order to identify a match.+This part illustrates the types of image transformations that LTU engine ​can handle in order to identify a match.
 Image transformations can be broadly divided into several groups: ​ Image transformations can be broadly divided into several groups: ​
-  ***[[functionalities#​geometric_transformations|Geometric transformations]]**: Includes scale changes, rotations, translations,​ flips and projective transformations. +  ***Geometric transformations**:​ Includes scale changes, rotations, translations,​ flips and projective transformations. 
-  ***[[functionalities#​photometric_transformations|Photometric transformations]]**: Includes color space conversions,​ gray level transformations,​ changes in hue, brightness and contrast. +  ***Photometric transformations**:​ Includes color space conversions,​ gray level transformations,​ changes in hue, brightness and contrast. 
-  ***[[functionalities#​image_filtering_and_noise|Filtering effects]]**: Includes noise and smoothing. +  ***Filtering effects**: Includes noise and smoothing. 
-  ***[[functionalities#​structural_transformations|Structural transformations]]**: Includes composite images, i.e. images that have been overlaid on top of each other, overlay of text, adding of borders and cropping. +  ***Structural transformations**:​ Includes composite images, i.e. images that have been overlaid on top of each other, overlay of text, adding of borders and cropping. 
-  ***[[functionalities#​compression_and_image_encoding|Recompression]]**: Due to different compression algorithms (such as jpg) and different image encodings, information can get lost and artifacts may appear.  +  ***Recompression**:​ Due to different compression algorithms (such as jpg) and different image encodings, information can get lost and artifacts may appear.  
-  ***[[functionalities#​images_derived_from_mobile_devices|Images derived from mobile]]**+  ***Images derived from mobile**
  
 Often images may be modified with a combination of the above transformations. However, the LTU image matching technology is robust even in those instances. Often images may be modified with a combination of the above transformations. However, the LTU image matching technology is robust even in those instances.
Line 99: Line 95:
 \\ \\
   * __Projective distortions__   * __Projective distortions__
-LTU Engine matching technology is capable of handling some degrees of perspective distortions. ​+Projective distortions appear when an object or a scene is pictured under different viewpoints, also referred to as perspective. ​LTU Engine matching technology is capable of handling some degrees of perspective distortions. ​
 \\ \\
 \\ \\
Line 128: Line 124:
 \\  \\ 
 \\ \\
-__Framed, flipped, text added, cropped__+__2.1.2.4.1 Framed, flipped, text added, cropped__
 \\ \\
 Also, LTU Engine matching technology is capable of handling: Also, LTU Engine matching technology is capable of handling:
Line 141: Line 137:
 \\ \\
  
-__Composite ​Images__+__2.1.2.4.2 Composite ​Images__
  
  
-A composite image contains several photographs or graphics in one image and often has a modified background or added text. For this kind of transformation,​ LTU Engine’s image matching technology delivers extremely accurate results. ​+A composite image contains several photographs or graphics in one image and often has a modified background or added text.  For this kind of transformation,​ LTU Engine’s image matching technology delivers extremely accurate results. ​
 \\  \\ 
 \\  \\ 
Line 168: Line 164:
 \\ \\
 \\ \\
 +
 +==== - Visual Distance ====
 +
 +
 +By using the DNA, LTU Engine computes a visual distance between the query image and the images in the reference database. Usually it is preferred to consider the order (ranking) in which the returned images are presented. \\
 +The visual distance is normalized such that a visual distance below 1.0 indicates a match. The distance is an indicator for the relevance of the retrieved images: the closer the value to 0.0, the closer the retrieved image shares the same visual content as the query image. Identical images have a distance of exactly 0.0. 
 +\\
 +\\
 +{{ :​imagealgorithms:​algo-001.jpg?​nolink&​1000 |}}
 +
 +\\ 
 +In addition, it is also possible to explore matches with a distance greater than one, i.e. matches with low confidence. As a result, the retrieval rate of matches that would otherwise not be included can be improved; however this also includes an increased rate of false positives. ​
 +\\ 
 +\\
 +{{ :​imagealgorithms:​algo-002.jpg?​nolink&​1000 |}}
 +
 +\\ 
 +The distance of this last example is greater than 1.0 as it only represents a fraction of the query image and thus only shares a small portion of the same visual content. Such images have a low rank and will thus figure last in the results, however will most likely be very relevant for many applications using the technology.  ​
 +
 +<​html>​
 +<​pagebreak>​
 +</​html>​
 +
  
 ==== - Matching Zone ==== ==== - Matching Zone ====