Presentation is loading. Please wait.

Presentation is loading. Please wait.

New horizons in the artificial vision

Similar presentations


Presentation on theme: "New horizons in the artificial vision"— Presentation transcript:

1 New horizons in the artificial vision
RETINA New horizons in the artificial vision 1-Introduzione 2-Caratteristiche 3-Caratteristiche funzionali 4-Case history 5-Conclusioni

2 It has the ability to learn and recognize
What is RETINA? RETINA is a C/C++ library based on artificial intelligence for digital image processing. It has the ability to learn and recognize objects in an image. RETINA 1 2 3 4

3 On what consists RETINA?
C/C++ library The library has a complete set of functions to manage all the features of RETINA. GUI configure models manage samples of the image run SVL analyze images RETINA 1 2 3 4

4 Retina Library Library files:
sb.dll (Windows) sb.so (Linux) sb.h It exports a complete set of functions to manage: image samples project file generic ROI images elaboration SVL RETINA 1 2 3 4

5 Retina GUI With the GUI you can train the models.
The result of the training is a project file. RETINA project file is all you need to process images. RETINA project file RETINA 1 2 3 4

6 Library integration Minimal integration:
develop code for images elaboration use GUI for learning Full integration: develop GUI for learning RETINA 1 2 3 4

7 Image elaboration 1 2 3 4 Load RETINA project file Load image
sb_retina_load() Load image sb_image_load() Create the ROI sb_roi_create() Elaborate image sb_retina_detection() RETINA 1 2 3 4

8 Never as easy as when you use the perception!
RETINA Segmentation? Never as easy as when you use the perception! RETINA 1 2 3 4

9 Why is it so difficult to achieve?
Segmentation It means identifying each object in the image . The human vision system can do it so well and without any effort of will that we believe it is an easy task to perform. Why is it so difficult to achieve? RETINA 1 2 3 4

10 Perception Do squares A and B have the same colours?
Then, our brain is deceiving us? RETINA 1 2 3 4

11 Perception Perception creates great confusion because:
what seems clear to the human eye is hidden to the image processing algorithms RETINA 1 2 3 4

12 «The perception is the expectation of finding a model»
What is the perception? When our brain sees an image, it always tries to simplify it as much as possible Our perception is always willing to organize what it sees in the most logical and comprehensible set. So you can say that: «The perception is the expectation of finding a model» RETINA 1 2 3 4

13 Perception versus Vision
For a vision system A and B have the same color. For a visual perception system A and B do not have the same color. The most correct definition of RETINA? «A visual perception system» RETINA 1 2 3 4

14 RETINA Characteristics RETINA 1 2 3 4

15 Characteristics Generic analysis not dedicated to any specific task
No configuration parameters needed It learns through the training Supervised learning (SVL) with human-machine interaction Multi-models management Scale management Collaborating/competing models management Models perturbations management Synthesis of multiple shapes in one model without losing the details Tolerant to: perspective distortion, blur, light change, image contrast, noise, shape deformation Support for multi thread and multi core processing RETINA 1 2 3 4

16 How to train RETINA RETINA is trained and ready to be used
The operator creates a set of images, which is representative of models variability The operator manually selects at least one sample for each model The operator starts the interactive SVL procedure RETINA is trained and ready to be used RETINA 1 2 3 4

17 What is the SVL? . . . it learns! You teach . . . You are the teacher
RETINA is the schoolboy . . . it learns! You teach . . . RETINA 1 2 3 4

18 The operator defines the target: the SVL works in order to achieve it.
The operator’s goal is to have an application which find objects in the images. The target is represented by the objects that the operator selects in the images. In technical language it is called “ground truth”. The SVL uses the "ground truth" and the background to extract and classify information in order to self-organize knowledge so that the "ground truth" is differentiated from the background. RETINA 1 2 3 4

19 System requirements Library for x86 architecture
Porting on other architectures and embedded systems is possible Does not require dedicated hardware (no GPU) SO: Windows, Linux Balances RAM/HD according to data base size RETINA 1 2 3 4

20 Field Applications Industrial (e.g. object detection, object classification, defects recognition) Robot vision Automation Quality control Sorting machines Security (e.g. pedestrian, people counter) Video Surveillance AV Image Processing Automotive (e.g. pedestrian, traffic signal recognition) ADAS – Advanced Driver Assistance Systems Intelligent Rearview Monitoring Systems Driver Monitoring System Autonomous Driving Traffic (e.g. vehicle classification, vehicle counting) Toll collection Congestion detection Traffic monitoring Parking Management RETINA 1 2 3 4

21 Functional characteristics
RETINA Functional characteristics RETINA 1 2 3 4

22 An eye to details but tolerant when needed
RETINA 1 2 3 4

23 Tolerant to perspective distortion
RETINA 1 2 3 4

24 Tolerant to blur RETINA 1 2 3 4

25 Tolerant to low image contrast
This is a red fish RETINA 1 2 3 4

26 Tolerant to noise RETINA 1 2 3 4

27 Tolerant to shape deformation
RETINA 1 2 3 4

28 Robust to partial occlusions
RETINA 1 2 3 4

29 Shadow? No problem! RETINA 1 2 3 4

30 Tolerant to light change
Fish is moving from sun to shadow RETINA 1 2 3 4

31 Rotation on 360° RETINA 1 2 3 4

32 Do you think it’s impossible?
RETINA Do you think it’s impossible? Let me try! RETINA 1 2 3 4

33 Beyond limits of vision
RETINA Beyond limits of vision


Download ppt "New horizons in the artificial vision"

Similar presentations


Ads by Google